Sodium Channel Splicing in Heart Failure Trial

Scientific Background and Significance

Introduction:

Congestive heart failure (CHF) represents a major health care concern in the United States. It has been estimated that approximately 5 million patients in the U.S. have CHF, and nearly 550,000 people are diagnosed with this disease annually.1 It is known that sudden cardiac death occurs more frequently in the setting of structural heart disease. Moreover, the risk for sudden cardiac death is 6 to 9 times greater in the heart failure population, and cardiac arrhythmias are perhaps the leading cause of death in CHF patients. 2,3 Currently, both the American College of Cardiology and the American Heart Association endorse the placement of internal cardioverter-defibrillators (ICDs) in patients with ischemic cardiomyopathy, reasonable life expectancy, and reduced ejection fraction below 40% (class I, level of evidence A).4 Additionally, placement of ICDs is recommended in non-ischemic cardiomyopathy patients who meet similar requirements with an ejection fraction of less than 35% (class I, level of evidence B).4 Despite these recommendations for primary prevention of sudden death by way of ICD implantation, more than half of the patients receiving a device are likely to not experience an arrhythmic event that necessitates ICD shock delivery.3 ICD devices, on average, cost $20-50,000 exclusive of operative and follow up costs. Currently, risk stratification of sudden cardiac death and the need for ICD placement is essentially dependent upon assessment of left ventricular ejection fraction. Other methods employed for risk stratification are signal averaged electrocardiogram (ECG) and another electrocardiographic technique known as T-wave alternans. Although these methods are FDA approved from risk prediction of cardiac death, such techniques are not widely employed in the U.S. given equipment and personnel costs to implement them. Thus, alternative testing for risk assessment for the development of sudden cardiac death in the heart failure population is desirable.

Role of Sodium Channels and the SCN5A Gene:

The cardiac voltage-gated sodium (Na+) channel, SCN5A, is the main channel generating current for electrical propagation in heart muscle and is the target of many antiarrhythmic drugs. Defective expression of the cardiac Na+ channel results in increased arrhythmic risk as evidenced by sudden death in the Brugada Syndrome.5 SCN5A mutations have also been implicated in the inherited long-QT syndrome, which can result in the development of the fatal dysrhythmias like ventricular fibrillation and torsades de pointes.6 Additionally, mutations in the SCN5A gene have also been proposed to exist and enhance risk for drug-induced dysrhythmias.7 Many studies have been done to shed light on the role of this tetrodotoxin-insensitive sodium channel in disease states. It has been demonstrated that mutated sodium channels in dilated cardiomyopathy may function differently depending upon the specific mutation type of the principal Na+ channel alpha-subunit.8 Specifically, Nguyen et al have demonstrated that these mutations may lead to changes in physiological function such as slower action potential rise time, enhanced late sodium current during steady state, or impaired inactivation.8 Additional mutations in the SCN5A gene have been linked to shifts in voltage dependence of Na+ channel inactivation in patients with idiopathic ventricular fibrillation.9 Additional research has concluded that decreased inactivation of late sodium currents may contribute to action potential prolongation.10 A different SCN5A gene abnormality, the E161K mutation, has been shown to lead to decreased sodium current density and a an 11.9 mV positive shift in the cell membrane half-maximal activation potential.11 Therefore, mutations of the Na+ channel can cause altered channel behavior and arrhythmias.

Valdivia et al. demonstrated that peak sodium current density is reduced 39% in the dog model, and approximately 57% in explanted failing human hearts, although the mechanism is unclear.10 Though our understanding of electrophysiological changes in heart failure and their relationship to arrhythmogenesis remains unclear, it can be inferred that disruptions in sodium handling - and perhaps increased intracellular sodium - may also result in change in calcium homeostasis via action of the Na+/Ca2+ exchanger.12,13 Overall, such changes in INa are likely to significantly contribute to arrhythmia in the setting of failing myocardium.

At the genome level, research has focused on the role of SCN5A gene mutations in arrhythmogenesis.8,9,14-16 Nevertheless, we have recently described acquired defects in Na+ channel messenger RNA (mRNA) that result in reduced Na+ current and occur only in failing hearts. Three 3'-terminal SCN5A mRNA splicing variants were identified and characterized in failing human heart ventricles. These splice variants were predicted to result in the translation of nonfunctional sodium ion channels that lack an appropriate domain IV pore region. These three splicing variants for the nonfunctional sodium channel gene product were denoted E28B, E28C, and E28D. Relative levels of the full length mRNA isoform, E28A, were decreased by 24.7% in patients with CHF compared to control. Additionally, two truncated versions of the gene product were increased in heart failure patients, reflecting an acquired genetic abnormality in production of the SCN5A gene in heart failure. At the same time, the E28C and E28D mRNA abundances were increased 14.2 fold and 3.8 fold respectively in CHF patients compared to controls. In order to confirm that truncation variants could contribute to arrhythmic risk, a gene-target mouse model was created with a nonsense mutation at exon 28. The electrophysiological effect of the presence of this gene truncation was also examined. It was found that action potential rate of rise was reduced (p=0.02, n=11), and the action potential amplitude was reduced from 76 ± 1.4 mV to 52 ± 0.6mV (p≤0.01, n=11). Integrated effect of this truncation was studied by examining bipolar field potentials (FP) of cardiomyocytes on microelectrode arrays.17 A 70.5% reduction (p<0.05) from -1126 ± 314 μV (n=6) in WT to 332 ± 174 micro V (n=7) in FP amplitude was observed. Additional measurements done by this method revealed a delay in FP rise and slowed conduction velocity in the mutant cells versus normal controls, suggesting that the truncation mutants could cause electrical abnormalities severe enough to contribute to arrhythmic risk. Also, we showed that lymphocytes process sodium channels similarly to cardiomyocytes. Thus, lymphocyte SCN5A mRNA processing may serve as a surrogate marker to assess SCN5A at the cardiac level and may correlated with arrhythmic risk in high risk populations. This study will assess that assertion.

Specific Aims/Objectives:

Research Aim 1:

To determine the abundances of SCN5A mRNA splice variants in patients with CHF and baseline ejection fractions less than 35% versus normal controls of similar age groups.

Research Aim 2:

To compare the abundances of SCN5A mRNA splice variants in patients with ICD devices who have and have not experienced ICD shock therapy.

Performance Sites of Research:

University of Illinois Medical Center and Jesse Brown VA Medical Center.

Key Research Personnel:

The Principal Investigator for this study will be Dr. Samuel Dudley, MD, PhD. Dr. Dudley is the Chief of the Section of Cardiology at the University of Illinois at Chicago. He is a Professor of Medicine and a published author in the field of cardiovascular medicine and physiology.

Research Methods:

Study Design:

Research Aim 1: We will correlate the amount of white cell Na+ channel splice variants with ejection fraction in patients with an without heart failure.

Research Aim 2: We will correlate the amount of white cell Na+ channel splice variants with the number of appropriate ICD shock in patients with ICDs in place.

Study Interventions:

This study requires no change in the standard of care. All study participants will be subjected to phlebotomy at the time of enrollment.

Sample Collection and Processing:

About 15 ML of blood will be drawn from study participants from University of Illinois at Chicago (UIC) or Jesse Brown Veterans Affairs Medical Center (JBVAMC) who have given informed consent for phlebotomy and study participation. Samples will be delivered immediately by study staff (coordinator, CO-PI, etc) to Dr. Dudley's (PI) lab in Room 1133 of the Clinical Sciences Building (CSB) for processing within 2 hours of collection. Levels of mRNA will be measured and some of the processed sample may be stored in a -80° F freezer in the same Lab for up to 7 years. Samples will NOT be stored or processed at JBVAMC or any other facility.

Statistical considerations:

The relationship of Na+ channel mRNA variant abundances will be compared in subjects with and without heart failure and in subjects with and without ICD events. The primary endpoint will be the a comparison of mRNA variant abundances. Dependent variables will include heart failure for aim 1 and number of shocks in aim 2. The number of patients need for each aim is determined by the variance of the test, the mean difference expected, and some consideration of the number of covariates that will need to analyzed in the regression analysis. Previously, we showed that the least sensitive measure was a reduction in E28A abundance by 24%. If we assume that the same percentage reduction will happen in aims 1 and 2, then we would need about 45 patients in each group to have a 90% power to detect this difference, assuming a 10% loss rate due to technical errors in the assays. Therefore, we would need a total of 180 patients for the total trial, 45 with heart failure, 45 controls, 45 ICD patients with events, and 45 patients without events.

Baseline data will be expressed as mean ± standard deviation (SD) for continuous variables, and frequencies for categorical variables. Differences in baseline characteristics between the groups will be examined by use of Fisher exact and Mann-Whitney tests for categorical and continuous variables, respectively. Because the number of ICD events recorded is a function of the observation time, Poisson regression will be used to model any relationship. In this model, the number of ICD events observed is assumed to be distributed following a Poisson distribution. That is, for a given period of time, the probability that a certain number of events has occurred is a function of the event rate multiplied by the duration of observation. In order to estimate the effects of mRNA variants on the rate of event occurrence, it is assumed that the event rate is log linear with respect to the predictors of interest. Solving this equation gives rate ratios comparing the rate of event occurrence in subjects with and without ICD events. Multiple expressions for the mRNA variant abundances will be considered, such as the relative abundance of each variant individually, the abundance of the individual variant as a function of the total Na+ channel mRNA, and the ratio of the truncations to the full-length Na+ channel mRNA. The regression coefficient will be estimated for the relationship between the dependent variable, ICD events, and the independent variables as the log of the rate ratio estimates. Statistical significance will be determined by using the likelihood ratio test. A p-value of 0.05 or less will be taken to be statistically significant. Results will be reported as the risk ratio and its associated 95% confidence interval. In order to select variables to be included in the model, we will consider, conservatively, those variables with a different distribution between the two groups at a p<0.20. The possibility of multicolinearity will be evaluated. Linear and non-linear terms will be considered. Normality of the variable distributions will be tested by a normal probability plot and by a Shapiro-Wilk test. While regression is fairly tolerant of violations in this regard, transformations will be investigated as necessary. Homoscedasticity will be evaluated by plot of residuals versus predicted values. Discrimination of the model will be evaluated by an overall C index and validated by bootstrap methods.

Anticipated Results and Pitfalls:

We do not anticipate any complications with acquiring blood samples, analyzing mRNA variants, or performing the statistics. Nevertheless, the major limitation to this trial is its retrospective nature. Nevertheless, this data will be useful in the design of future prospective trials.