PITX2 Modulates Atrial Membrane Potential and the Antiarrhythmic Effects of Sodium-Channel Blockers

Background Antiarrhythmic drugs are widely used to treat patients with atrial fibrillation (AF), but the mechanisms conveying their variable effectiveness are not known. Recent data suggested that paired like homeodomain-2 transcription factor (PITX2) might play an important role in regulating gene expression and electrical function of the adult left atrium (LA). Objectives After determining LA PITX2 expression in AF patients requiring rhythm control therapy, the authors assessed the effects of Pitx2c on LA electrophysiology and the effect of antiarrhythmic drugs. Methods LA PITX2 messenger ribonucleic acid (mRNA) levels were measured in 95 patients undergoing thoracoscopic AF ablation. The effects of flecainide, a sodium (Na+)-channel blocker, and d,l-sotalol, a potassium channel blocker, were studied in littermate mice with normal and reduced Pitx2c mRNA by electrophysiological study, optical mapping, and patch clamp studies. PITX2-dependent mechanisms of antiarrhythmic drug action were studied in human embryonic kidney (HEK) cells expressing human Na channels and by modeling human action potentials. Results Flecainide 1 μmol/l was more effective in suppressing atrial arrhythmias in atria with reduced Pitx2c mRNA levels (Pitx2c+/–). Resting membrane potential was more depolarized in Pitx2c+/– atria, and TWIK-related acid-sensitive K+ channel 2 (TASK-2) gene and protein expression were decreased. This resulted in enhanced post-repolarization refractoriness and more effective Na-channel inhibition. Defined holding potentials eliminated differences in flecainide’s effects between wild-type and Pitx2c+/– atrial cardiomyocytes. More positive holding potentials replicated the increased effectiveness of flecainide in blocking human Nav1.5 channels in HEK293 cells. Computer modeling reproduced an enhanced effectiveness of Na-channel block when resting membrane potential was slightly depolarized. Conclusions PITX2 mRNA modulates atrial resting membrane potential and thereby alters the effectiveness of Na-channel blockers. PITX2 and ion channels regulating the resting membrane potential may provide novel targets for antiarrhythmic drug development and companion therapeutics in AF.

A trial fibrillation (AF) causes cardiovascular death, frequent hospitalization, and cognitive decline even in patients treated according to guidelines (1-3). Antiarrhythmic drug (AAD) therapy remains the most commonly used treatment to maintain sinus rhythm in AF patients, but AAD effectiveness remains limited (3). Unfortunately, we lack a basic understanding of why AADs prevent AF over long periods in some patients but not in others (4,5). Identifying factors that modify the effects of AADs would allow the selection of responsive patients and could help guide development of novel AADs (6).
Paired like homeodomain-2 transcription factor (PITX2) is a transcription factor that regulates the development of the left atrium (LA) and thoracic organs. Its c isoform is expressed in the adult LA and regulates the expression of LA ion channels (7)(8)(9). Low atrial Pitx2 expression renders mice susceptible to AF and shortens the LA action potential (8,10,11). In this study, we investigated how atrial PITX2 modifies the effects of AADs.
We detected variable LA PITX2 messenger ribonucleic acid (mRNA) expression in AF patients requiring rhythm control therapy. After finding that low Pitx2c enhanced the effect of flecainide, mediated by a more positive resting membrane potential (RMP), we identified reduced TWIK-related acid-sensitive K þ channel 2 (TASK-2) expression as a possible driver of this effect and replicated these effects in cells expressing human sodium (Na) channels and in a human atrial action potential model. Single nucleotide polymorphisms (SNPs) rs2200733, rs6838973, and rs1448818 (14) were identified using TaqMan assays (Thermo Fisher Scientific Inc., Waltham, Massachusetts).

All experiments were conducted under the Animals
Adult mice (age 12 to 16 weeks) on an MF1 background with normal or reduced (Pitx2c þ/À ) atrial Pitx2c expression were studied (8).
The human atrial cell model of Courtemanche et al. (19) was used. Pitx2c þ/deficiency was modeled by reducing I K1 conductance by 25% and doubling I Kr conductance. Simulations were run in strands of 100 atrial cells (cell length 100 mm). The 5 leftmost cells of the strand were paced (S1) for 2 min at 1,000-and 500-ms basic cycle lengths. Premature stimulation (S2) was applied to determine the ERP and conduc-   refractoriness (PRR) was calculated as the difference between APD at -60 mV repolarization and ERP.
Ribonucleic acid and complementary deoxyribonucleic acid were synthesized from murine LA, (SuperScript VILO, Thermo Fisher Scientific Inc.) to quantify expression of 20 atrial ion channels and genes with suspected PITX2-dependent regulation (9) using custom-designed Taqman low density array plates (Thermo Fisher Scientific Inc.). Western immunoblotting was performed on murine LA tissue lysates with antibodies detecting TASK-2, K v 1.6, Na/K ATPase alpha-1, Na/K ATPase alpha-2, Na/Ca exchanger 1, Serca2a, Na v 1.5, or calnexin, using standard methods.
Optical action potentials and calcium ion (Ca 2þ ) transients were recorded in murine LA and analyzed using custom-made MATLAB algorithms (MathWorks, Natick, Massachusetts) as previously described (17).

RESULTS
PITX2 mRNA varied markedly in human LAA (Central Illustration) harvested from AF patients (   Values are mean AE SD (range), n, or mean (range). *Left atrial appendages were collected from these patients with atrial fibrillation (AF).
PVI ¼ pulmonary vein isolation.  [13]). Although PITX2 mRNA is numerically lower in patients with 5 or 6 risk alleles, we did not find a PITX2 mRNA gradient according to AF risk.
Kcna6 and Kcnk5 mRNA expression were reduced in Pitx2c þ/murine LA ( Figure 4A, Online Table 1), whereas mRNA concentrations of 20 other ion channels or related genes were not altered. Kv1.6 protein concentration was unaltered, whereas TASK-2 protein concentration was reduced in murine atria with reduced Pitx2c expression ( Figure 4B). Na v 1.5 mRNA  Figures 5D and 5E).
Background K þ currents, which include TASK currents, were reduced in Pitx2c þ/murine atria, whereas I K1 did not differ between genotypes ( Figure 6).  Table 6), consistent with published data (8). We found that 1 mmol/l flecainide decreased atrial conduction velocities without differences between wild-type and Pitx2c þ/mice (Online Figures 1B and 1C). Calcium transient relaxation times at 50% relaxation were not different between wild-type and Pitx2c þ/-(Online Figures 1D   and 1E). Flecainide 1 mmol/l shortened 50% Ca 2þ relaxation times by approximately 10% and decreased Ca 2þ transient amplitude by approximately 50% in murine atria with normal and reduced Pitx2c expression (Online Figures 1E and 1F). Additionally, expression of the Na/Ca exchanger Serca2a and Na/K ATPase alpha-1 and alpha-2 subunit protein did not differ between wild-type and Pitx2c þ/atria (Online Figure 2).

DISCUSSION
This study demonstrated that LA PITX2 mRNA con-  Syeda et al. PITX2 expression is confined to the LA in the heart, AAD therapy that leverages modifications in RMP may achieve "atrial-specific" AAD therapy.
RMP is maintained by an intricate balance of different transmembrane currents and is closely related to the potassium equilibrium potential. We identified that PITX2 modifies expression of the genes encoding K v 1.6 and TASK-2 ( Figure 4). Complete deletion of PITX2 regulates other potassium and Na channels such as Kcnj2 (8,36), which alter the RMP, but these were not responsible for the depolarized RMP observed in our study. Two-pore domain potassium channels, such as TASK-2, contribute to RMP in various cells, including skeletal and cardiac muscle (37,38). To date, an altered function of the TASK-1 channel and of I K1 has been implicated in atrial remodeling and AF (39,40). This study demonstrated that TASK-2 is expressed in atrial myocardium ( Figure 4B), suggesting that a reduced function of TASK-2 could depolarize RMP (Figures 1 and 5) (8,11), analogous to the effect of TASK-2 in neuronal and cartilage tissue (41,42).

DEVELOPING CLINICAL MARKERS FOR PATIENTS
WITH DEPOLARIZED RMP. It will be challenging to directly assess LA RMP in AF patients, but our data suggested that differences in atrial RMP could explain the effectiveness of Na-channel blockers in carriers of   an assumption that has not been definitively proven (9,11,44,46). Our analysis (   Values are mean AE SEM (number of atria).
-¼ not applicable; other abbreviations as in Table 3.   The Task  Syeda et al.