IJBEM logo
International Journal of Bioelectromagnetism
Vol. 4, No. 2, pp. 79-80, 2002.

previous paper

next paper



Y.  Rudy
Cardiac Bioelectricity Research and Training Center, 509 Wickenden Bldg.,
Case Western Reserve University, Cleveland, Ohio 44106-7207, USA

Abstract: Dispersion of myocardial repolarization plays a major role in the initiation and maintenance of arrhythmias.  Increasingly, seemingly idiopathic syndromes that lead to cardiac arrhythmias and sudden death  are being linked to genetic defects.  In my presentation, I will illustrate how computational biology can be used to link such genetic defects to repolarization abnormalities of cardiac cells.  I will also demonstrate that an ECG imaging modality (ECGI) can detect and locate regions of  increased dispersion of  repolarization in the heart.  Work to be described has been published previously [1-5].


Abnormalities of myocardial repolarization and in particular dispersion of repolarization, underlie many life threatening arrhythmias. In recent years, many such repolarization disorders have been linked to genetic defects in a particular ion channel. An example is the long-QT (LQT) syndrome, which is associated with a prolonged Q-T interval on the ECG and increased incidence of syncope and sudden cardiac death. While various clinical phenotypes of LQT have been linked to genotypes, establishing a mechanistic link between genotype and phenotype remains a major challenge.  Since action potential generation involves complex interactions between various ion channels, other membrane transport processes, and the dynamically changing ionic environment of the cell,  it is extremely difficult to predict the cellular  electrophysiological consequences of a particular channel defect. Computational biology and mathematical modeling is one approach that can provide mechanistic insights in this context.  In the first part of my presentation I will illustrate how computational  biology  can be used to bridge the gap between a particular channel defect and the resulting repolarization abnormality in LQT.

Despite the role of dispersion of  repolarization in arrhythmogenesis, current noninvasive methods for detecting  such dispersion from the  body surface ECG have shown severe inconsistencies and limitations. In contrast, epicardial measures of repolarization can provide accurate information on the existence , location, and severity of dispersion of repolarization in the heart. In the second part of my presentation,  I will describe results of applying a noninvasive ECG imaging modality (ECGI) to reconstruct epicardial measures of dispersion of repolarization from body surface ECG data.


1. LQT Simulations [1-3].  

The Luo-Rudy dynamic (LRd) model of the cardiac ventricular cell was the basis for the simulations. The model represents all important membrane ion channels, pumps and exchangers. It accounts for the dynamic concentration changes of Na+, Ca2+, and K+ during the action potential for  pacing at various rates. State specific (Markov) models of the wild type (WT) and mutant ion channels were developed and validated. The channel models were then incorporated in the LRd cell model and the action potential  was computed at various rates of pacing.

2. ECGI Imaging of Repolarization [4, 5].

Epicardial potentials were recorded with a 224-electrode sock from an open chest dog during control, regional left ventricular (LV) warming, LV cooling, and simultaneous adjacent LV warming and cooling to induce localized changes in myocardial repolarization and regions of increased dispersion. Body surface potentials were forward computed from the recorded epicardial potentials in a human torso model containing the lungs, sternum, spine, and skeletal muscle. Realistic geometrical errors and measurement noise were added to the torso data to simulate data acquired in the clinical setting. Epicardial potentials were then computed from the body surface potentials using Tikhonov second order regularization. Measured epicardial data served as gold standard for validating noninvasively reconstructed potentials, electrograms, and derived measures of dispersion of repolarization ( activation–recovery intervals, ARI,  and QRST integrals).


1. LQT Simulations.

LQT3: A Sodium Channel Mutation [1, 3].     Several distinct mutations in SCN5A, the gene that encodes the cardiac sodium channel , give rise to LQT3. The most severe is the DKPQ mutation, a deletion mutation of three amino acids from a region of the channel protein that participates in inactivation of the channel. This structural defect results in two modifications of channel function: (1) secondary channel reopenings following the first opening and (2) transient failure of inactivation leading to channel bursting. The Markov model of the channel simulates correctly these two types of  kinetic behavior. When introduced into the whole-cell model, the mutant channels generate a late (sustained) current during the action potential (AP) plateau which acts to prolong  the AP duration. AP prolongation is more severe at slow pacing rates, with arrhythmogenic early afterdepolarizations (EADs) developing at cycle length > 1000 ms. This observation  is consistent with the clinical presentation of LQT3 where arrhythmias occur during bradycardia (sleep).The presence of 50% mutant channels is sufficient to cause EADs. The mechanism of EAD formation is recovery from inactivation and reactivation of L-type calcium channels during the prolonged AP plateau.

LQT2: A Potassium Channel Mutation [2].     Mutations in the HERG gene, the gene that encodes the major subunit of IKr , lead to LQT2.  Using a similar strategy to the LQT3 study above, we developed single-channel Markovian models of wild-type and mutant IKr , and introduced them into the LRd ventricular cell model .  More than 50 mutations in HERG have been linked to LQT2.  We chose to study several well-characterized mutations that affect the channel kinetics (T474I and R56Q), cause loss of ion selectivity (N629D), or exert their effects through a negative dominant mechanism or trafficking defects.  Important observations include : (1) The severity of the phenotype depends on the specific kinetic changes and how they affect IKr  during the time-course of the AP.  (2) R56Q speeds IKr deactivation, which decreases the current late in the AP causing significant AP prolongation.  (3) T474I affects the current predominantly early in the AP, causing only minor AP prolongation.  (4) N629D, by shifting the reversal potential of IKr to –13mV, results in an inward current (carried by Na+ ions) during the late phase of the AP plateau, causing AP prolongation and EADs.  (5) All mutations, including those with loss of function due to dominant-negative effects or trafficking defects, prolong AP duration much more in mid-myocardial cells than in epicardial cells.  This can increase dispersion of repolarization and create conditions for the development of reentrant arrhythmias.

2. Imaging Dispersion of Repolarization [4, 5].

Noninvasively reconstructed epicardial electrograms captured faithfully the T-wave changes induced by warming and cooling. During regional LV warming, the reconstructed electrograms over the warmed region showed augmented positive T-waves compared to control. During cooling, electrograms showed inverted and negative T-waves over the cooled region. During simultaneous warming and cooling, positive and negative T-waves were reconstructed noninvasively over the warmed and cooled regions, respectively.   ARIs over the warmed region decreased during warming, reflecting local shortening of AP duration due to accelerated kinetics of repolarizing currents. During cooling, ARIs increased over the cooled region, reflecting prolonged AP duration due to slowed (by the cooling) kinetics of repolarizing currents. Steep local ARI gradients were generated during simultaneous adjacent LV warming and cooling These changes were faithfully captured in the noninvasive ECGI reconstructions.   Similarly, epicardial QRST integral maps clearly reflected and located the regional repolarization changes induced by the warming/cooling interventions. Importantly, these changes were not reflected in the body surface QRST integral maps, but were faithfully captured in the noninvasively reconstructed epicardial QRST maps using ECGI.


The LQT simulations demonstrate the complexity of the excitatory process even at the level of a single cardiac cell.  The cellular behavior is determined by complex interactions between various cellular processes. For example, in LQT3 INa   mutation underlies AP prolongation, while in LQT2 it is caused by IKr.    In both cases, arrhythmogenic EADs are generated by the L-type calcium channels , an “innocent bystander” that is not affected by the mutations.  The mutations prolong the AP, providing sufficient time at the appropriate membrane potential for recovery and reactivation of the L-type channels   which , in turn , generate the EAD.

Body-surface ECG and potential maps do not provide reliable indices of the degree of dispersion of myocardial repolarization. Moreover, they cannot be used to locate regions  of  high dispersion (an arrhythmogenic substrate) in the heart. The study demonstrates that ECGI imaging has the capability to reconstruct repolarization properties and localize areas of increased dispersion of repolarization  in the heart noninvasively.  Clinically, ECGI may provide a useful tool for noninvasive risk stratification, and for guidance and evaluation of therapy of repolarization-related arrhythmias.

Acknowledgments:  Work supported by grants R37-HL33343 and  RO1-HL49054 from the National Institutes of Health – National Heart, Lung and Blood Institute.


The presentation will be based on the following publications:

[1] C. E. Clancy, Y. Rudy, "Linking a genetic defect to its cellular phenotype in a cardiac arrhythmia" Nature 400:566-569, 1999.

[2] C. E. Clancy, Y. Rudy, “Cellular Consequences of HERG Mutations in the Long QT Syndrome: Precursors to Sudden Cardiac Death” Cardiovascular Research   50:301-313, 2001.

[3] C.E. Clancy, Y. Rudy, “A Na+ Channel Mutation that Causes Both Brugada and Long QT Syndrome Phenotypes: A Simulation Study of Mechanism” Circulation  2002 (in press)

[4] J.E. Burnes, R.N. Ghanem, A.L. Waldo, Y. Rudy, “ Imaging Dispersion of Myocardial Repolarization I. Comparison of Body Surface and Epicardial Measures”   Circulation  104:1299 – 1305, 2001.

[5] R.N. Ghanem, J.E. Burnes A.L. Waldo, Y. Rudy, “ Imaging Dispersion of Myocardial Repolarization II. Noninvasive Reconstruction of  Epicardial Measures”   Circulation 104:1306 – 1312, 2001.


previous paper table of contents next paper

© International Society for Bioelectromagnetism