Decreased Left Atrial Cardiomyocyte Fibroblast Growth Factor 13 Expression Increases Vulnerability to Postoperative Atrial Fibrillation in Humans

Postoperative atrial fibrillation (POAF) complicates 30% to 40% of cardiac surgeries and increases short‐ and long‐term morbidity and mortality.¹ Advancement in developing novel POAF therapeutics has been prevented by an inadequate understanding of molecular mechanisms promoting vulnerability to atrial fibrillation (AF).¹ In this observational cohort study, we performed transcriptome analyses in left atrial myocytes from patients undergoing mitral valve surgery and identified differential gene expression associated with new‐onset POAF. Notably, the most significant association with POAF was decreased expression of fibroblast growth factor 13 (FGF13),² the depletion of which we demonstrate is sufficient to prolong the action potential in rat atrial myocytes.

After institutional review board approval and informed consent, we enrolled adult patients without a history of AF presenting for elective mitral valve surgery for severe degenerative mitral regurgitation. Patients with prior cardiac surgery, mitral stenosis, aortic regurgitation, aortic stenosis, and endocarditis were excluded. Patients were consecutively enrolled at Ronald Reagan UCLA Medical Center between April 2021 and June 2022. POAF was defined as any clinician diagnosis of AF by ECG of at least 30‐seconds duration from intensive care unit admission until hospital discharge. All patients had continuous ECG monitoring throughout their hospital course, which was independently reviewed daily for validation of POAF diagnosis. Patients did not receive prophylactic amiodarone. Among 28 patients enrolled with no history of AF (Figure [A]), 15 experienced POAF (53.6%), an incidence consistent with the increased risk of POAF after mitral valve surgery.¹ Preoperative corrected QT prolongation was the only clinical characteristic significantly associated with POAF (P=0.04, t test). Separately, we enrolled 2 patients with preexisting persistent AF presenting for the same surgery for qualitative comparison of POAF‐associated gene expression.

During cardiopulmonary bypass, the surgeon obtained a small (≈0.1‐g) sample of left atrial tissue along the standard incision on the left atrium at Waterston's groove for surgical exposure of the mitral valve. Cardiomyocytes were isolated from frozen human left atrial tissue samples, as previously described.³ RNA was extracted from cardiomyocytes using the Qiagen AllPrep Micro Kit. RNA libraries were sequenced, and read pairs were pseudoaligned to hg38. We used DESeq2 to identify differential gene expression in our transcriptome‐wide analysis (false discovery rate [FDR]<0.05). RNA‐sequencing analyses demonstrated enrichment in cardiomyocyte markers (Figure [B]). Differential gene expression analyses revealed 13 genes associated (FDR<0.05) with POAF (Figure [C]). Among these, FGF13 was the most significant (Figure [D]). The association of decreased FGF13 expression with POAF was highly significant (P=1.52x10⁻⁸, FDR=0.0001) after controlling for age, sex, and corrected QT. FGF13 splice variants were not identified in our RNA‐sequencing data set based on analyses with DEXseq. FGF13, a fibroblast growth factor homologous factor expressed in cardiomyocytes, is known to modulate voltage‐gated sodium channel Na_V1.5 inactivation.² Given the POAF incidence in our study, we were adequately powered (80%) to detect effect sizes ≥1.6 between groups (FDR‐adjusted 2‐tailed α=0.0022, assuming an FDR rate of 5% and a true‐positive rate of 5%) using the 2‐sample t‐test as a simplification of the DESeq2 negative binomial model. Thus, our RNA‐sequencing analysis is powered to detect large effect sizes with significant clinical relevance in our data set. The raw fastq files from the left atrial cardiomyocyte RNA‐sequencing data set will be available at the database of Genotypes and Phenotypes (dbGAP) (https://www.ncbi.nlm.nih.gov/gap/), accession number phs003644.v1.p1.

To isolate neonatal rat atrial myocytes, atria were dissected from euthanized postnatal day 3 Sprague‐Dawley rats and enzymatically digested. Atrial myocytes were isolated and plated on 24‐well plates. After 24 hours, myocytes were co‐transfected with Cy3‐conjugated non‐targeting siRNA and either FGF13‐targeted or negative control siRNA for 12 hours. After an additional 48 hours, transfected myocytes were identified by Cy3 fluorescence, followed by whole‐cell patch clamp experiments to measure action potential duration (APD) at 90% repolarization at 2‐second cycle length or quantitative reverse transcription–polymerase chain reaction/immunoblot to confirm FGF13 knockdown. Current clamp experiments demonstrated that FGF13 knockdown (Figure [E–G]) caused significant atrial myocyte APD prolongation (control siRNA median APD at 90% repolarization=0.110 seconds; interquartile range=0.081‐0.179 seconds, n=10; FGF13 siRNA median APD at 90% repolarization=0.217 seconds; interquartile range=0.142‐0.282 seconds, n=12; exact P=0.03, Wilcoxon rank‐sum test; Figure [H] and [I]).

FGF13 has been shown to be protective in inhibiting pathologic late sodium current,⁴ which has been implicated in AF pathogenesis by prolonging APD. Accordingly, FGF13 knockout was shown to increase the late sodium current.⁵ In this study, we provide critical insight into the relevance of FGF13 dysregulation to acquired arrhythmia in humans: (1) we measure FGF13 mRNA in human left atrial samples obtained in proximity to the pulmonary veins, the most common anatomical origin of AF; (2) we link decreased FGF13 levels to new‐onset POAF; (3) we show that FGF13 expression is decreased in patients with POAF before the onset of AF, providing a clinically actionable target for POAF prophylaxis; and (4) we demonstrate APD prolongation in atrial myocytes after FGF13 knockdown, validating a molecular mechanism for arrhythmia. Our findings support a model whereby decreased FGF13 expression creates vulnerability to AF through left atrial myocyte APD prolongation that, following the stress response to surgery, precipitates AF in the postoperative period.