Introduction
Nonvalvular atrial fibrillation (AF) is the most common sustained cardiac arrhythmia. AF results from electrophysiological abnormalities and alteration of atrial tissues, which promote the formation and propagation of abnormal electrical impulses. The prevalence of AF is gradually increasing worldwide, especially among older adults, and this is causing important health, social, and economic problems, and increases the risk of stroke, heart failure, morbidity, and disability [
1].
Diagnosis and prognosis in AF are based on evaluation of symptoms, monitoring of irregular rhythm, and assessment of cardiac anatomical structure and function; however, most AF patients are asymptomatic and effective long-term monitoring is challenging. Therefore, the identification of blood-derived biomarkers may help physicians evaluate AF risk and severity.
The development and propagation of AF are influenced by electrical and structural remodeling of the atria. Predisposing factors associated with electrophysiological abnormalities in AF are aging, cardiovascular risk factors (e.g., hypertension, diabetes, dyslipidemia, heart failure, obstructive sleep apnea, and obesity), genetic mutation, and dysregulation of ion channels and transporter expression [
2]. Current studies have focused a great deal of attention on gene regulatory mechanisms, particularly microRNA [
3]. MicroRNAs (miRNAs) are small non-coding ribonucleic acid (RNA) molecules that average 22 nucleotides in length that mediate post-transcriptional gene regulation by sequence-specific inhibition of target mRNA translation. They are essential in a variety of biological processes, including proliferation, differentiation, apoptosis, and metabolism [
4]. More than 2500 miRNAs have been identified in the human genome, and more than 1500 miRNAs have had their gene regulatory functions formally defined [
5]. MicroRNAs that are released into extracellular fluids (e.g., plasma, urine, saliva), are called circulating miRNAs or extracellular miRNAs. Three pathways of miRNA secretion have been reported, including incorporation with high-density lipoprotein (HDL), binding with Argonaut 2 (Ago2), and encapsulation in extracellular vesicles [
6].
Extracellular vesicles (EVs) are lipid bound vesicles that are released by many cell types into the extracellular space. EVs are found in various body fluids such as blood, urine, cerebrospinal fluid, and saliva. Increasing evidence suggests that EVs act as a vehicle to transfer genetic materials in cell-to-cell communication [
7]. MicroRNAs derived from extracellular vesicles (EV-miRNAs) play important roles in both normal homeostasis and pathophysiology in many diseases, including cardiovascular diseases (CVDs) [
8,
9]. EV-miRNAs have been described as a selective packaging mechanism that parental cells sort out a set of miRNAs into EVs for secretion to target cells [
10,
11]. International Society for Extracellular Vesicles (ISEV) has suggested to classify EVs based on difference in size as small EVs (< 100 nm or < 200 nm) and medium/large EVs (> 200 nm) [
12]. Small EVs (sEVs), also referred to as exosomes, are smallest types of EVs (30–150 nm) and released by inward budding of multivesicular bodies (MVBs) from plasma membrane into the extracellular space. Large EVs (lEVs), also called microvesicles (MVs) or microparticles (MPs), are secreted by outward budding from plasma membrane during cell activation or cell stress and have a size ranging from 100 to 1000 nm.
Expression of EV-miRNAs has been shown to be associated with AF and reflect pathophysiology of AF [
13‐
15]. However, all previous studies focused on sEV-miRNAs, there are no evidences of the relationship between lEV-miRNAs and AF. lEV-miRNAs have been reported to be diagnostic and prognostic biomarkers in several diseases. Jansen F, et al. reported that increased expression of miR-126 and miR-199a in MPs, but not freely circulating miRNA expression, predicts the occurrence of cardiovascular events in patients with stable coronary artery disease [
16]. miR-129-5p isolated from plasma MVs was demonstrated to be a sensitive and specific biomarker for heart failure (HF) in univentricular heart disease [
17]. In addition, MP miR-124a and miR-150, which were found to be more abundant in obese subjects compared to normal weight controls, were found to be significantly associated with inflammation and vascular function in obesity [
18]. As a consequence, lEV-miRNAs are emerging as attractive biomarkers due to their stability and easy detection in biofluids. lEV-miRNA detection can be sensitive, predictive, specific, and noninvasive, which are all characteristics of an ideal biomarker [
19]. The aim of this study was to investigate for differences in lEV-miRNA expression profiles between AF patients and non-AF controls. Subsequently, lEV- miRNAs of interest were validated using droplet digital polymerase chain reaction (ddPCR), and bioinformatics analysis was used to predict miRNA target genes and their functional pathways. The results of this study revealed that the expression signature of six lEV-miRNAs that reflect pathophysiology of AF.
Material and methods
Study population and subject enrollment
AF patients were recruited from the Division of Cardiology, Department of Medicine and the control subjects were recruited from Department of Preventive and Social Medicine, Faculty of Medicine Siriraj Hospital, Bangkok, Thailand during November 2019 to December 2020. The protocol for this study was approved by the Siriraj Institutional Review Board (SIRB) (COA no. Si 489/2019), and complied with the principles set forth in the Declaration of Helsinki (1964) and all of its subsequent amendments. All study participants provided written informed consent prior to inclusion. Forty-two AF patients (3 new-onset AF, 28 paroxysmal AF, 8 persistent AF, and 3 permanent AF) and 42 age- and sex-matched non-AF controls were enrolled. All participants underwent a thorough historical investigation and 12-lead electrocardiography (ECG) to confirm cardiac rhythm. Twelve AF patients and 12 non-AF controls were enrolled for the discovery phase, and the other 30 AF patients and 30 non-AF controls were included in the validation phase. Non-AF controls had no history of CVDs, no history of atrial arrhythmias, and no current cardiovascular treatment. No study participants had history of myeloproliferative disorders, thrombocytopenia, ischemic stroke within the previous 3 months, malignancy, rheumatic mitral valve disease, acute infectious or inflammatory disease, or pregnancy.
Blood processing and lEV isolation
Peripheral blood that was collected into BD Vacutainer® K2EDTA Plus Blood collection tubes (BD Biosciences, Franklin Lakes, NJ, USA), was centrifuged at 1500 ×
g for 15 min to isolate platelet poor plasma (PPP), and stored at − 80 °C until analysis. Blood specimens were processed within 4 h after blood draw. lEV isolation was performed as previously described [
20]. Briefly, PPP was centrifuged at 17,000 ×
g for 2 min at 4 °C to remove remaining platelets, and the lEVs were pelleted by centrifugation at 17,000 ×
g for 45 min at 4 °C. lEV pellets were washed with filtered phosphate-buffered saline (PBS) prior to resuspension with 100 μl of fresh PBS.
Nanoparticle tracking analysis (NTA)
lEV concentration and size distribution were measured using a NanoSight NS300 (Malvern Panalytical, Malvern, UK) equipped with a 488 nm laser. lEV suspension was diluted (1:100–1:200) in filtered PBS. Samples were analyzed under constant flow conditions (flow rate: 30) at 25 °C, and were captured with a camera level of 13–14 using NTA software version 3.4 (Malvern Panalytical). Five independent measurements (60 s each) were obtained for each sample. Data are reported as mean ± standard deviation (SD).
Transmission electron microscopy (TEM)
lEV suspensions were fixed in 2% glutaraldehyde (Sigma-Aldrich Corporation, St. Louis, MO, USA) in PBS for 30 min at 4 °C, and then absorbed onto 200 mesh copper grids with carbon-coated formvar film for 15 min. The grids were negatively stained with 2% uranyl acetate (w/v) for 3 min, and then the excess liquid was removed by blotting with filter paper. Grids were imaged under a transmission electron microscope (JEM-1230; JEOL Ltd, Tokyo, Japan) at 100 kV.
Flow cytometry
Five microliters of each lEV suspension was analyzed using a CytoFLEX S flow cytometer (Beckman Coulter Life Sciences, Indianapolis, IN, USA) and CytoExpert analysis software (Beckman Coulter Life Sciences). lEV gate was indicated by size calibration beads (Spherotech, Inc., Lake Forest, IL, USA) between 100 and 1300 nm. lEV sizes were analyzed using 405 nm violet side scatter. Data analysis was performed using FlowJo software (version 10 for Windows) (FlowJo, LLC, Ashland, OR, USA).
Western blot analysis
The amounts of total protein from lEV samples were determined with a bicinchoninic acid assay kit (Pierce, Thermo Scientific, Rockford, IL, USA), according to the manufacturer’s instructions. Western blot analysis was used to determine the EV protein markers, cardiomyocyte marker and EV purity in lEV samples. All samples were adjusted to 30 μg of total protein before mixing with reducing sample buffer and loaded into 10% SDS–polyacrylamide gel electrophoresis. Then, proteins were transferred to a PVDF membrane (Amersham Hybond PVDF Membrane, GE Healthcare Life Sciences, Freiburg, Germany). Membrane was blocked with 5% w/v non-fat milk in TBST buffer and then incubated with primary antibodies; anti-CD63, anti-Alix, anti-Apolipoprotein A (Abcam, Cambridge, MA, USA) and anti-Caveolin 3 (Invitrogen, Thermo Fisher Scientific, Waltham, MA USA) overnight at 4 °C. After washing, membranes were incubated with HRP linked goat anti-rabbit IgG (Abcam Cambridge, MA, USA) for 1 h at room temperature. Chemiluminescent detection was performed using Clarity Western ECL Substrate (Biorad Laboratories, Inc, Hercules, CA, USA). Protein bands were visualized by ImageQuant LAS 4000 (GE Healthcare Life Sciences).
Quantitative reverse transcription polymerase chain reaction (RT-qPCR) array
Twelve AF patients and 12 non-AF controls were pooled into 4 samples in each group (3 subjects per 1 sample). Total lEV-RNA was extracted using TRIzol LS® Reagent (Life Technologies, Carlsbad, CA, USA) following the manufacturer’s protocol. RNA extraction efficiency was checked via three spike-in controls: UniSp2, UniSp4, and UniSp5. Total RNA was quantified using a Qubit™ RNA HS Assay Kit and a Qubit® 2.0 Fluorometer (both Life Technologies). Total RNA and small RNA profiles were investigated using an Agilent 2100 Bioanalyzer system with an Agilent RNA 6000 Pico Kit and an Agilent Small RNA Kit, respectively (Agilent Technologies, Santa Clara, CA, USA). Reverse transcription was performed using a miRCURY LNA™ Universal RT microRNA PCR (Qiagen, Hilden, Germany) using cel-miR-39-3p and UniSp6 as internal control. Expression of lEV- miRNA profiles was measured using miRCURY LNA miRNA Serum/Plasma Focus PCR Panels (Qiagen) containing the 179 most abundant miRNAs in circulation. RT-qPCR was performed using a CFX96™ Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA) with a miRCURY LNA™ SYBR Green PCR Kit (Qiagen). All samples passed internal quality control (QC) checks for extraction, reverse transcription, qPCR efficiency, and hemolysis. The data analysis was performed on the QIAGEN web portal at the GeneGlobe Data Analysis Center (
https://geneglobe.qiagen.com/us/analyze/). The results were reported as cycle threshold (Ct) values, which were normalized using the global Ct mean of expressed miRNAs method. lEV-miRNA expression was calculated as fold change relative to non-AF controls using the 2
−ΔΔCT method.
Droplet digital polymerase chain reaction (ddPCR)
In the validation phase, lEV-miRNA expression was confirmed in 30 AF patients and 30 non-AF controls using a ddPCR™ system (Bio-Rad Laboratories). lEV-RNA concentration was measured using a NanoDrop™ 8000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Two hundred nanograms of total RNA was reverse transcribed using a miRCURY LNA™ Universal RT microRNA PCR (Qiagen). One microliter of synthesized cDNA was added to a 20 μl PCR reaction mixture containing 10 μl of 2× EvaGreen Supermix (Bio-Rad Laboratories), 1 μl miRCURY LNA miRNA PCR primer (Qiagen), and 8 μl RNase-free H
2O. Twenty microliters of ddPCR assay mixture was loaded into a disposable droplet generator cartridge (Bio-Rad Laboratories) with 70 μl of QX200 Droplet Generation Oil for EvaGreen. Detailed information about primers for miRNAs is presented in Additional file
1: Table S1. The cartridges were placed inside the QX200 Droplet Generator (Bio-Rad Laboratories), and then the droplet mixtures were transferred to a ddPCR™ 96-well PCR plate (Bio-Rad Laboratories). PCR amplification was performed using a C1000 Touch Thermal Cycler (Bio-Rad Laboratories). The thermal cycling conditions were, as follows: 95 °C for 5 min, 40 cycles of 95 °C for 30 s, and 54 °C for 1 min (ramping rate reduced to 2%), and three final steps at 4 °C for 5 min, 90 °C for 5 min, and 4 °C indefinite hold. A no template control (NTC) was included in every assay. The PCR-positive and negative droplets were read using a Q×200 Droplet Reader (Bio-Rad Laboratories). QuantaSoft software (Bio-Rad Laboratories) was used to quantitate the concentration of miRNAs, and the results are presented as the number of copies per microliter (no. copies/μl) of PCR reaction. Synthetic miRNA (Cel-miR-39-3p) was added into all samples to check RNA extraction efficiency.
MicroRNA regulation was analyzed using miRNA target gene prediction and pathway analysis. Diana-Tarbase version 8.0 (
https://carolina.imis.athena-innovation.gr/diana_tools/web), which is a database of experimentally supported miRNA-gene interactions, was used to identify the target gene of the miRNAs miR-106b-3p, miR-590-5p, miR-339-3p, miR-378-3p, and miR-532-3p. The TargetScan miRNA target prediction tool (
http://www.targetscan.org/vert_72/) was used to predict target genes of miR-328-3p. Prediction of miRNA function and pathway was performed using Diana-miRPath version 3 (
http://snf-515788.vm.okeanos.grnet.gr/) with default settings. The miRNA target genes were categorized according to biological process, molecular function, and cellular component using the Gene Ontology (GO) (
http://geneontology.org/) bioinformatics database. Functional pathways related to the cardiovascular system were identified by Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways analysis. The threshold value was set at
p < 0.05.
Statistical analysis
Data were analyzed using PASW Statistics version 18 (SPSS, Inc., Chicago, IL, USA). Baseline characteristics compared between two groups was analyzed using chi-square test for categorical variables, and those results are reported as number and percentage. Unpaired t-test and Mann–Whitney U test were used to compare normally distributed (reported as mean ± standard deviation [SD]) and non-normally distributed continuous variables (reported as median and interquartile range [IQR]), respectively. Multivariate logistic regression analysis was performed to investigate for association between miRNA levels and AF. The results of logistic regression analysis are reported as odds ratio (OR) and 95% confidence interval (CI). Significant difference was defined as a p-value less than 0.05.
Discussion
This is the first study of lEV-miRNA profile in AF patients, and our results suggest several important implications. First, we found upregulation of 6 lEV-miRNAs to be significantly associated with AF and to reflect important pathophysiology of AF. Second, the expression of selected lEV-miRNAs was evaluated by ddPCR, which is a highly sensitive and accurate technique that sidesteps the need to use unsuitable reference genes, and that eliminates false positive and false negative results. This technique may assist in the development of lEV-miRNA platform for clinical application in the future.
Many studies have reported a relationship between circulating miRNAs and AF. Downregulation of miR-150 was shown to be associated with AF in several studies [
21‐
24]. Decreased levels of plasma and atrial tissue miR-29b have been found in AF patients, which suggests its role in atrial fibrotic remodeling by targeting collagen-1A1 (COL1A1) [
25]. Low level of whole blood miR-328 was reported to be associated with prevalent AF [
22], whereas high level of plasma miR-328 was found in AF patients, especially plasma from the left atrial appendage [
26]. Due to the inconsistent results and unrelated levels of miRNAs between plasma and tissue, it has been suggested that circulating miRNAs may not useful biomarkers for AF [
23]. In contrast, extracellular vesicle (EV)-derived miRNAs are potentially attractive biomarkers since they are more specific and reflect the pathogenesis of AF. Upregulation of exosomal miRNAs (miR-103a, miR-107, miR-320d, miR-486, and let-7b) has been demonstrated in patients with persistent AF compared with supraventricular tachycardia controls, which suggests that these miRNAs are involved in atrial function and structure, oxidative stress, and fibrosis pathways [
14]. Wei, et al. found the expression levels of exosomal miR-92b-3p, miR-1306-5p, and let-7b-3p to be significantly increased in AF patients compared with normal sinus rhythm [
13]. Another study showed exosomal miRNAs (miR-438-5p, miR-142-5p, miR-223-3p, and miR-223-5p) to be related to AF. Multivariate logistic analysis suggests that miR-483-5p is independently correlated with AF [
15]. Epicardial fat-derived EVs (eFat-EVs) from patients with AF have been shown the upregulation of profibrotic miR-46b, whereas antifibrotic miR-33a and 29a are downregulated, compared to those without AF, which are associated with the stimulation of extensive myocardial fibrosis in the heart rats [
27].
Fibrillating atria in AF has been shown to promote more myolysis, nuclear alteration, and apoptosis, which related to activation of programmed cell death via strongly upregulated CASP-3 expression and downregulated BCL-2 expression [
28,
29]. Cell activation or apoptosis contributes to increased membrane permeability and remodeling, which subsequently results in lEV generation [
30]. Several mechanisms in AF have been described as potent inducers of apoptotic cell death, such as high ventricular heart rate, low or oscillatory shear stress, stretch, hypoxia, inflammation, and oxidative stress, and all of these mechanisms can promote lEV generation [
31]. Therefore, we hypothesized that the content within lEVs, especially miRNAs, may relate to the pathophysiology of AF.
There were no significant differences in total number, mean size and mode size of lEVs between AF patients and non-AF controls, suggesting that it may result from non-EV protein contamination. The relationship between lEV size and concentration was deeply investigated, which concentration of lEVs at the size of 101–200 nm in AF patients was significantly higher than non-AF controls. Elevated levels of this size range may result from the pathophysiology of AF that stimulates the mechanism of lEV generation [
32]. For example, acute induction of AF activates platelets within minutes and significantly increases the expression of P-selectin on both platelets and platelet-derived lEVs [
33]. The increase of circulating procoagulant lEVs might reflect a hypercoagulable state that could contribute to atrial thrombosis and thromboembolism. Mechanical stretch has been linked to Ca
2 + overload in cardiomyocytes in AF [
31,
34]. Increasing of intracellular Ca
2+ activates cytoskeleton cleavage through calpain and caspase activation, leading to membrane remodeling ultimately results in lEVs shedding. Our previous study showed that circulating lEVs are released from platelets, endothelial cells, leukocytes and red blood cells [
35]. However, more than 45% of lEVs in AF patients were not able to be characterized their cellular origins. In this study showed increased expression of Caveolin-3 in lEV samples from AF patients. These results revealed that cardiomyocytes are one of the sources of lEVs that may relate to an increase in lEV level at the size of 101–200 nm in AF patients.
lEV-miRNA profiling analysis revealed 19 significantly upregulated miRNAs, and 21 significantly downregulated miRNAs in AF patients compared to non-AF controls. The levels of six highly expressed miRNAs (miR-339-3p, miR-106b-3p, miR-378a-3p, miR-590-5p, miR-328-3p, and miR-532-3p) were confirmed with absolute quantification in our validation cohort, and logistic regression analysis showed the elevated levels of these miRNAs to be significantly associated with AF. Moreover, bioinformatics analysis demonstrated these 6 highly expressed miRNAs to likely be involved in arrhythmogenesis, cell apoptosis, cell proliferation, oxygen hemostasis, and structural remodeling—all of which are processes implicated in the pathogenesis of AF.
An abundance of evidence suggests that miRNAs may be directly or indirectly involved in AF by modulating atrial electrical remodeling (miR-328-3p, miR-106b-3p) and structural remodeling (miR-590-5p). miR-328-3p regulates particular genes involved in inflammation, myocyte depolarization (CACNA1C and CACNB1), vascular function (ABCG2), and cellular aging (H2AFX) [
36‐
39]. Elevation of miR-328 level in left atrial tissue has been found in both canine AF models and AF patients with rheumatic heart disease. Overexpression of miR-328 in a canine model of AF diminished L-type Ca
2 + current and shortened atrial action potential duration – both of which increase vulnerability to AF [
40]. Previous reports have been suggested that many miRNAs promote atrial structural remodeling and fibrosis, including miR-328, miR-31, miR-1, miR-146b and miR-21, they are also detected in the circulation of various cancer patients. These miRNAs have been hypothesized that may relate to the high incidence of AF in cancer patients via their repressive effects in the genes that control cardiac arrhythmias [
41].
Chiang et al. reported the miR-106b-25 cluster (miR-25, miR-93, and miR-106b) to be mediators of electrical remodeling, and downregulation of the miR-106b-25 cluster has been found in the atria of patients with AF. Loss of the miR-106b-25 cluster leads to upregulation of ryanodine receptor type-2 (RyR2) protein levels and pro-arrhythmic sarcoplasmic reticulum Ca
2 + -release, which are associated with increased AF susceptibility [
42]. Nicotine use is associated with downregulation of miR-590-5p, which partly explains the upregulation of TGF-β1 and TGF-βR2 proteins in the right atrial appendage of human and canine AF models. These two proteins promote the production of collagens in the myocardium, which was found to relate to the development of myocardial fibrosis, and the subsequent induction of atrial structural remodeling and fibrillation [
43].
miR-378a-3p, which is the most abundant miRNA in skeletal muscles, including cardiac muscles, was reported to be involved in anti-apoptosis [
44], angiogenesis [
45], anti-fibrosis [
46], and anti-hypertrophy [
47]. Downregulation of miR-378a-3p has been found in the right atrial appendages (RAA) and left atrial appendages (LAA) in AF patients with rheumatic mitral valve disease [
48]. It is, therefore, possible that downregulation of miR-378a-3p partly promotes cardiac apoptosis, fibrosis, and hypertrophy in AF. Two studies reported the pro-apoptotic role of miR-532-3p via suppression of apoptosis repressor with caspase recruitment domain (ARC) [
49,
50]. Upregulation of whole blood miR-339-3p has been observed in coronary heart disease [
51]. Ming Tana, et al. reported that the levels of miR-339-3p increase in mouse plasma exosomes before thrombosis, and that they are enriched in thrombin-stimulated platelet-derived exosomes in vitro, which suggests association with platelet activation [
52]. Many studies have shown high levels of platelet-derived MPs in AF compared to healthy controls, which suggests their potential involvement in platelet activation [
35,
53,
54]. In the present study, the level of lEV-derived miR-339-3p was upregulated in AF patients, which may result from selective packaging of miR-339-3p from activated platelets into their lEVs during AF.
Upregulation of 6 lEV-miRNAs in AF patients may be involved in two main mechanisms. First, lEVs act as a mediator in cell–cell communication. Upregulation of miR-378a-3p, miR-590-5p, and miR-106b-3p in lEVs may reflect cardioprotective signaling for anti-arrhythmia, anti-fibrosis, anti-apoptosis and anti-hypertrophy, which means that upregulation of these lEV-miRNAs could attenuate the pathogenesis of AF. Second, upregulation of miR-328-3p, miR-532-3p, and miR-339-3p may be involved in the generation of lEVs following cardiomyocyte activation or apoptosis, which could release high levels of apoptosis-related miRNAs in lEVs into blood circulation.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.