Replicated gene expression changes in patients with atrial fibrillation

Funding Acknowledgements

Type of funding sources: Public grant(s) – EU funding. Main funding source(s): European Union Horizon 2020 CATCH ME; Cardiovascular Research Netherlands RACE V

Background

Little is known about changes in the atrial transcriptome associated with paroxysmal and persistent atrial fibrillation (AF).

Purpose

To identify major molecular mechanisms in AF, we determined consistent differential expression (DE) between atrial tissue samples from well-characterized patients with paroxysmal or persistent AF and patients without a history of AF (no AF) in two independent patient cohorts.

Methods

Poly-A tailed RNA from left and right atrial appendage tissue samples from independent discovery and replication cohorts CATCH ME (n=192) and RACE V (n=122) was sequenced and analyzed according to patient AF history. Analyses were performed stratified by atrial side, adjusting for age, sex, heart failure and a combination of clinical characteristics determined by principal component analysis. Transcripts were considered DE in CATCH ME if their fold change reached transcriptome-wide significance (false discovery rate (FDR) < 0.05). DE transcripts in each rhythm comparison were replicated in RACE V if we observed a concordant direction of effect and a within-set FDR < 0.05 in the same comparison.

Results

Persistent AF compared to no AF was associated with 184 left atrial DE transcripts in CATCH ME of which 85 (46%) were replicated in RACE V, and with 208 right atrial DE transcripts in CATCH ME of which 86 (41%) were replicated in RACE V. Overall, 26 transcripts were discovered and replicated in both atria. Discovered but non-replicated transcripts often did exhibit concordant direction of effect (left: 78%, right: 83%). Replicated transcripts consisted of protein coding genes, antisense and non-coding RNAs. Protein coding genes showed involvement in pathways linking persistent AF to cardiomyocyte structure, conduction properties, fibrosis, inflammation, molecule trafficking, and endothelial dysfunction. Interestingly, paroxysmal AF was not consistently associated with DE transcripts in any comparison. Principal component analysis of the expression of the 26 transcripts strongly associated with persistent AF did however reveal a distinct paroxysmal AF expression profile in-between no AF and persistent AF patients in the first principal component scores (Figure 1).

Conclusion

RNA sequencing of human atrial tissue samples identified many transcripts associated with persistent AF in left and/or right atria, discovered and replicated using two independent cohorts. These consistent findings of AF-induced changes provide a starting point for targeted proteomic analysis and single-nucleus sequencing to further unravel the molecular mechanisms underlying AF progression to persistent AF, and biomarker development to quantify AF progression and enable precision medicine in individual patients.

RNA-seq profiling of the atrial transcriptome reveals gender-specific patterns and interactions with atrial fibrillation and heart failure

Funding Acknowledgements

Type of funding sources: Public grant(s) – EU funding. Main funding source(s): TRAIN-HEART Innovative Training Network, funded by the European Union’s Horizon 2020 research and innovation program (under the Marie Sklodowska-Curie grant agreement no. 813716) Characterizing Atrial fibrillation by Translating its Causes into Health Modifiers in the Elderly (CATCH ME), funded by the European Union’s Horizon 2020 research and innovation program (under the grant agreement no. 633196)

Background

Atrial fibrillation (AF) is the most common cardiac arrhythmia and is associated with heart failure (HF) and stroke. Clinical and experimental data from previous studies suggest gender differences in mechanisms and phenotypes of AF: women may have more atrial fibrosis, worse outcomes after catheter ablation, and some women carry a higher risk for thromboembolic complications than men. The molecular mechanisms underlying these differences are still poorly understood.

Methods

Gender-based transcriptional patterns were assessed using paired-end, directional RNA sequencing data generated from atrial tissue biopsies in 199 patients either in sinus rhythm or with paroxysmal or persistent AF as part of the CATCH-ME project. Transcript counts were compared between genders separately in the left and right atria using the DESeq2 package in R. The models were adjusted for potential sources of confounding (age, atrial fibrillation status, heart failure status and sequencing batch). Interaction models were implemented using DESeq2 to compare gender*morbidity interactions for persistent AF and HF. Significance was assessed using likelihood ratio tests comparing models with and without the interaction terms. Results with an adjusted P-value  0.05 were considered significant and utilized for subsequent downstream assessments. Differentially expressed (DE) genes were tested for enrichment of gene ontology (GO) terms and KEGG pathways using the WebGestalt toolkit.

Results

Transcriptome-wide profiling across the cohort identified 33 sex-differentiated genes in the left atria and 51 in the right atrial samples, with 21 of these showing bilateral differences. Interestingly, 36 (44%) of the results from these analyses were comprised of non-coding transcripts, including long non-coding RNAs (lncRNAs), antisense RNAs and pseudogenes. GO and pathway enrichment analyses for these genes revealed their involvement in critical pathways such as the complement and coagulation cascades and RNA transport.  Interaction analyses between gender and AF identified two genes (MPP2 & GNAS-AS1) that were differentially transcribed in the right atria and one gene (MYL2) that was DE in the left atria by gender in persistent AF samples. A similar analysis comparing gender*HF morbidity also revealed evidence of DE. Four transcripts (HLA-DQB1-AS1, EIF1AY, UTY and ZFY-AS1) showed gender-specific differences in expression by HF status in left atria, while HLA-DQB1-AS1 was differentially regulated by gender and HF status in right atrial samples.

Conclusions

These RNA-seq analyses provide novel insights into gender-related differences in the transcriptional landscape of right and left adult human atrial appendages. Moreover, interaction analyses identified three genes DE in female atria in persistent AF and four DE genes in female atria in heart failure, providing a molecular anchor for the observed differences in atrial diseases phenotypes between men and women.

Telomere length is associated with atrial fibrillation

Background

Atrial Fibrillation (AF) is common and is caused and predisposed to by a complex pathophysiology. Aging is among the most important risk factors for AF. Yet, some individuals develop AF in early years, whereas elderly individuals may remain free of AF. Aside from measurable concomitant risks, we hypothesized that a pathophysiologically relevant biological age exists, which outweighs a patient's calendar age. Telomere length is a measurable marker of age, which might reflect biological age. AF and telomere length have previously been associated, but results remained controversial. Here, we tested the relation between AF and telomere length in a well-characterized and so far largest cohort.

Methods

Since 2005, we enrolled 2475 patients with AF from the prospective AFLMU cohort, preferentially if they developed AF before age 65 years, and 3077 control individuals free of AF from the community-based KORA Study between 2006–08. All participants received a detailed clinical characterization, an electrocardiogram, and a blood draw for biomarker analyses. In all participants, we determined telomere length using a qPCR-based method. In a 384 well format, we employed a multiplex TaqMan assay to determine both telomere length and the single copy gene 36B4. Telomere length was expressed by the delta-CT method and was reformatted to have lower CT values indicate shorter telomere length. We compared telomere length between cases and controls using multi-variably corrected logistic regression models.

Results

Our cohort's mean age was 58 years in AFLMU and 56 years in KORA F4. Men were enrolled more commonly, with 72.3% in AFLMU and 51.7% in KORA F4. For consistency with available information, we confirmed that telomere length is continuously decreasing with age and that men have shorter telomere length compared to women. As a main result we found that AF patients have significantly shorter telomere length compared to controls (controls: telomere length 13,10 [12.60, 13.63] versus AF: 9.81 [5.98, 13.1], p<0,001). This relation remained significant following multi-variable adjustment for sex, body mass index, hypertension, and diabetes. The odds ratio per unit longer telomer length was 0.77 [95% confidence interval 0.74–0.77], p<0.001. We also calculated a propensity-score matched model of cases with and without AF confirming our main results (controls: telomere length 13.21 [11.65, 13.67] versus AF: 9.03 [5.35, 12.54], p<0.001). The multi-variable adjusted model revealed an odds ratio of 0,75 [95% confidence interval 0.73–0.76], p<0.001.

Conclusion

AF is significantly associated with telomere length in one of the largest cohorts to date. Assessment of telomere length may adjudicate patients with AF due to premature biological aging. The underlying reasons for such premature aging remain to be identified.

Funding Acknowledgement

Type of funding source: Public grant(s) – EU funding. Main funding source(s): European Commission - Horizon 2020

P5662Development of a prognostic model for prevalent atrial fibrillation using individual patient data: Results of CATCH ME

Background/Introduction

Atrial fibrillation (AF) can be challenging to diagnose due to asymptomatic and paroxysmal presentation. Identifying prognostic factors of AF would elucidate potential mechanisms causing AF and refine screening for at risk patients.

Purpose

To identify the main predictors of AF and to develop a prognostic model for prevalent AF.

Methods

Data of 120 potential predictors were harmonised in individual patient data from 4 independent European studies. A three stage Delphi expert consensus process identified predictors based on clinical knowledge. The predictors were further reduced using statistical selection (backward elimination), and a logistic regression model was fitted. We calculated odds ratios (OR) for each of the selected predictors and evaluated model performance using the C-statistic.

Results

Overall, 2420 patients (mean [standard deviation] age = 62.7 [14.5] years, 35.6% female, 43.1% with AF) were included in the analysis. Thirty-one potential predictors identified from the Delphi process which had sufficient data across all datasets were modelled. Of these 14 were deemed prognostic in predicting AF (age, sex, BMI, height, hypertension, diabetes, history of coronary artery disease, left atrial volume, left ventricular end systolic diameter, abnormality on echo, tricuspid valve disease of at least moderate intensity, aldosterone-antagonists, beta-blockers and P2Y12 blockers; see Figure 1). There was a clear interaction between age and sex indicating that males are at higher risk than females early in life, while females are at increased risk of AF at older age (Figure 1). The risk prediction model combining these prognostic factors performed well (C-statistic 0.79; 95% CI 0.77–0.81).

Figure 1. (a) Forest plot; (b) Interaction

Conclusion(s)

Our preliminary analysis identified important prognostic factors and a complex relationship between age and sex, which predicts prevalent AF, highlighting the different potential causes of AF in different patients. There is a clear need to validate these factors in external datasets and for further investigation into the molecular mechanism underlying these factors.

Acknowledgement/Funding

European Commission H2020 framework

Screening for Atrial Fibrillation

Approximately 10% of ischemic strokes are associated with atrial fibrillation (AF) first diagnosed at the time of stroke. Detecting asymptomatic AF would provide an opportunity to prevent these strokes by instituting appropriate anticoagulation. The AF-SCREEN international collaboration was formed in September 2015 to promote discussion and research about AF screening as a strategy to reduce stroke and death and to provide advocacy for implementation of country-specific AF screening programs. During 2016, 60 expert members of AF-SCREEN, including physicians, nurses, allied health professionals, health economists, and patient advocates, were invited to prepare sections of a draft document. In August 2016, 51 members met in Rome to discuss the draft document and consider the key points arising from it using a Delphi process. These key points emphasize that screen-detected AF found at a single timepoint or by intermittent ECG recordings over 2 weeks is not a benign condition and, with additional stroke factors, carries sufficient risk of stroke to justify consideration of anticoagulation. With regard to the methods of mass screening, handheld ECG devices have the advantage of providing a verifiable ECG trace that guidelines require for AF diagnosis and would therefore be preferred as screening tools. Certain patient groups, such as those with recent embolic stroke of uncertain source (ESUS), require more intensive monitoring for AF. Settings for screening include various venues in both the community and the clinic, but they must be linked to a pathway for appropriate diagnosis and management for screening to be effective. It is recognized that health resources vary widely between countries and health systems, so the setting for AF screening should be both country- and health system-specific. Based on current knowledge, this white paper provides a strong case for AF screening now while recognizing that large randomized outcomes studies would be helpful to strengthen the evidence base.

Static and Dynamic Electrocardiographic Patterns Preceding Torsades de Pointes in the Acquired and Congenital Long QT Syndrome

The arrhythmogenic mechanisms involved in the triggering of the polymorphic ventricular tachycardia called torsades de pointes (TdPs) remains to be elucidated. In this work, we investigated the static and dynamic profiles of the repolarization interval from the surface electrocardiogram recorded in healthy individuals and in cardiac patients with TdPs. We implemented this analysis just prior to the arrhythmia onset and we computed the delta values based on baseline periods (1 hour prior to event). We measured QT/QTc prolongation, QT variability, ventricular ectopic beats (VPBs) frequency, T-wave amplitude, T-peak to T-end interval, and T-wave complexcity. The analysis of these parameters in reference to baseline revealed 1) an increased QTc variability, 2) the presence of VPCs, and 3) the profound changes in T-loop morphology in patients developing TdPs.

Acute respiratory distress syndrome with transiently impaired left ventricular function and Torsades de Pointes arrhythmia unmasking congenital long QT syndrome in a 25-yr-old woman

We report a case of recurrent episodes of Torsades de Pointes arrhythmia in the setting of transiently impaired left ventricular ejection fraction, acute respiratory distress syndrome, transient hypokalaemia and QT-prolonging drugs, in a previously healthy 25-yr-old female patient. In the course of the clinical and genetic work-up this patient was newly diagnosed with a mutation in KCNH2 encoding the alpha-subunit of the human repolarizing potassium channel I(Kr). This case report illustrates the multivariate nature of long-QT syndrome, and emphasizes the usefulness of a pharmacological test for repolarization abnormalities.