Abstract
Objective: Atrial fibrosis plays a critical role in atrial fibrillation (AF). A key event in the pathogenesis of fibrosis is the activation of fibroblasts (FBs) into myofibroblasts (MFBs). Paracrine factors released from MFBs lead to ion channel expression changes in cardiomyocytes (CMs). Downregulation of Ltype calcium channel Cav 1.2 expression is a hallmark of AFassociated ionic remodeling. However, whether exosome (Exo)mediated crosstalk between MFBs and CMs regulates Cav 1.2 expression remains unknown.
Methods: Atrial FBs and CMs were isolated and cultured from neonatal rats by enzymatic digestion. The activation of FBs into MFBs was induced by angiotensin II. Coculture assay and in vitro Exo treatment were used to determine the effect of MFBderived Exos on Cav 1.2 expression. Confocal Ca2+ imaging was performed to examine the adrenergic stimulationelicited Ca2+ influx signals. The levels of potential Cav 1.2inhibitory microRNAs (miRNAs) were measured by qRTPCR.
Results: Untreated FBs expressed limited amounts of alpha smooth muscle actin (αSMA), while angiotensin II induced a significant upregulation of αSMAexpressing MFBs. Cocultures of MFBs and CMs resulted in downregulation of Cav 1.2 expression in CMs,which was largely abolished by pretreatment of MFBs with exosomal inhibitor GW4869. More importantly, treatment with MFBderived Exos caused repression of Cav 1.2 expression in CMs. Additionally, the adrenergic receptoragonistelicited Ca2+ influx signals in CMs were remarkably attenuated by pretreatment with MFBderived Exos, corresponding to the paralleled change in Cav 1.2 expression. Finally, miR213p, a potential Cav 1.2inhibitory miRNA, was enriched in MFBderived Exos and upregulated in CMs in response to MFBderived Exos.
Conclusion: We uncover an Exomediated crosstalk between MFBs and CMs, contributing to increased vulnerability to AF by reducing the expression of Cav 1.2 in CMs.
Keywords
Exosome · Paracrine communication · Ltype calcium channel · Atrial fibrillation
Introduction
Atrial fibrillation (AF), the most common arrhythmia, contributes substantially to cardiac morbidity and mortality in China and the West [1]. The pathophysiological mechanisms underlying AF are complex. At the cellular level, the arrhythmogenic substrate is characterized by the abbreviation of action potential duration and effective refractoryperiod [2]. Decreased Ltype Ca2+ current (ICaL) and reduced protein expression of voltagegated Ltype calcium channel subunit α1c (Cav 1.2; encoded by Cacna1c) have been implicated in AF and are primarily responsible for the shortened action potential duration [3–6]. Although the reduced protein expression of Cav 1.2 is believed to be triggered by high atrial rate during AF [7], marked decreases in ICaL density and protein content of Cav 1.2 have already been observed in atrial cardiomyocytes (CMs) from patients with conditions that predispose to AF [8, 9].
Notably, atrial fibrosis plays a critical role in the occurrence and development of AF [10, 11]. A key event in the pathogenesis of myocardial fibrosis is the activation of fibroblasts (FBs) into myofibroblasts (MFBs), characterized by enhanced expression of alpha smooth muscle actin (αSMA) [12]. Growing evidence suggests that paracrine interactions between MFBs and CMs have the potential to favor the development of AF substrate. Coculture of CMs with MFBs or exposure of CMs to MFBconditioned medium results in abnormal ion channel expression, decreased Ca2+ transient amplitude, and shortened action potential duration [13–15]. However, the identity of the specific paracrine factors involved in electrophysiological changes remains largely unknown.
Recently the exosomes (Exos), a subtype of small extracellular vesicles, have garnered considerable attention for their potent role in paracrine signaling. Exomediated crosstalk between MFBs and CMs has been implicated in the development of CM hypertrophy [16, 17]. Nevertheless, to our knowledge, evidence that an Exorelated regulatory mechanism check details is involved in arrhythmia has not yet been directly addressed. This study demonstrates for the first time that Exomediated crosstalk between MFBs and CMs contributes to increased vulnerability to AF by reducing the expression of Cav 1.2 in CMs.
Materials and Methods
Cell Culture and Treatments
Atrial FBs and CMs were isolated and cultured from 1to 3dayold SpragueDawley rats by enzymatic digestion as described previously with slight modifications (see online suppl. methods; for all online suppl. material, see www.karger.com/ doi/10.1159/000505641). The high purity ofFB and CM cell cultures is shown in online supplementary Figure 1. Cells were cultivated in DMEM/F12 supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin at 37 ° C in 5% CO2. To promote theMFB phenotype, confluent monolayers of FBs were starved for 24 h in 0.1% Exodepleted fetal bovine serum and then cultured in DMEM/F12 supplemented with 0.1% Exodepleted fetal bovine serum with Ang II (100 nM) for 24 h. Expression of αSMA, fibronectin, connective tissue growth factor (CTGF), collagen type Ia1 (Col Ia1), collagen type IIIa1 (Col IIIa1), and osteopontin were evaluated by Western blot.
Immunofluorescence Staining
Cells were seeded onto polyLlysinecoated glass slides. After treatment as indicated, the cells were subjected to immunofluorescence staining as previously described [16]. Primary antibodies are showed in the online supplementary methods.
CoCulture Assay
Coculture assay was performed using transwell plates (membrane pore=0.4 μm). Atrial FBs and CMs were seeded in the upper and lower chamber, respectively, and were cultivated separately before the coculture experiment. Ang II was used to induce FB activation into MFBs. After washing with PBS, untreated FBs and Ang IItreated MFBs were cocultured with CMs in DMEM/F12 containing 0.1% Exodepleted fetal bovine serum for 24 h. To inhibit Exo production, untreated FBs and Ang IItreated MFBs were pretreated with GW4869 (10 μM) for 24 h, and then cells were cocultured with CMsinDMEM/F12 containing GW4869 (10 μM) for another 24 h.
Exo Isolation and Characterization
Exos were isolated from conditioned medium of untreated FBs and Ang IItreated MFBs cultured in DMEM/F12 supplemented with 0.1% Exodepleted fetal bovine serum. Conditioned medium was collected for 24 h, and Exos were isolated by several centrifugation and filtration steps (see online suppl methods) as described previously [16]. The Exo pellet was resuspended in PBS. Expression of CD63, tumor susceptibility gene 101 protein (TSG101), calnexin, and constitutive heat shock protein (HSC70) were analyzed by Western blot analysis. Exo samples were used for transmission electron microscopic analysis.
Measurement of Intracellular Free Ca2+ Concentration
Atrial CMs were grown on glass bottom dishes and pretreated with Exos for 24 h. After loading with fluo4/AM (5 μM) for 40 min in the dark at 37 ° C, cells were washed with Hanks’ balanced salt solution and then incubated in DMEM containing thapsigargin (1 μM),a potent irreversible inhibitor of sarcoplasmic reticulum (SR) Ca2+ATPase [18], for 30 min to deplete SR Ca2+ stores. Suppression of SR Ca2+ uptake allowed us to record Ca2+ influx signals without contamination by Ca2+ release from SR [18]. Thus, the intracellular free Ca2+ concentration ([Ca2+]i) rising exclusively from Ca2+ influx via voltagegated calcium channels (the predominant Cav channel form expressed in CMs is the Cav 1.2 channel [19]). The changes in [Ca2+]i were represented by relative fluorescence intensity compared with the average initial value before drug application (F/F0). Isoproterenol (ISO; βadrenergicagonist) and norepinephrine (NE; α,βadrenergicagonist) were used to elicit Ca2+ influx signals.
Western Blot Analysis
Protein expression was assessed by Western blot as described previously [17]. Primary antibodies are showed in the online supplementary methods.
RNA Isolation and qRTPCR Assay
Total RNAs were extracted from cells or Exos using TRIzol reagent. Reverse transcription reactions were conducted on samples including total RNA, RT primer, MMLV Reverse Transcriptase (Epicentre), 10× Reaction Buffer (Epicentre), RNase inhibitor (Epicentre), and dNTP (HyTest Ltd.). For realtime PCR, 2× Master Mix (Arraystar) was used. Data were analyzed using the 2–ΔΔCt method and normalized to levels of U6 (cellular) or miR93 (exosomal). The reaction protocol and primers sequences are available in the online supplementary methods.
Statistical Analysis
Data are presented as mean ± SEM. Data analyses were performed using SPSS 19.0. Ap value <0.05 was considered statistically significant. Student’s t test or oneway ANOVA, followed by SNK posttest, was applied to compare the data between two or among experimental groups, respectively.
Results
AngII Induces Activation of FBs intoMFBs
To create an in vitro model that mimicked a fibrotic cellular milieu, Ang II, a wellcharacterized profibrotic molecule, was added to atrial FBs. MFB activation was examined by αSMA expression. Untreated FBs expressed limited amounts of αSMA, while Ang II induced a significant upregulation of αSMAexpressing MFBs. Furthermore, Ang II increased fibronectin, CTGF, Col Ia1, Structural systems biology Col IIIa1, and osteopontin protein levels (Fig. 1).
CoCultures of MFBs with CMs Lead to Downregulation of Cav 1.2 Protein in CMs
To determine the potential effect of paracrine factors released by MFBs on Cav 1.2 expression, we first performed a coculture assay (Fig. 2a). Atrial CMs showed markedly reduced Cav 1.2 expression levels when cocultured with Ang IItreated MFBs, as compared with when cocultured with untreated FBs, indicating that MFBs produced soluble mediators that regulated Cav 1.2 expression in CMs. Furthermore, MFBinduced change in Cav 1.2 expression was significantly abolished by pretreatment of MFBs with exosomal inhibitor GW4869. The results suggested that this regulatory effect of MFBs was likely to be Exodependent.
Exos Derived from MFBs Are Major Paracrine Factors Mediating Cav 1.2 Downregulation in Atrial CMs
To further confirm the Exorelated regulatory mechanism, we then performed an in vitro Exo treatment assay. Exos were prepared from conditioned medium of untreated FBs and Ang IItreated MFBs. Transmission electron microscopic analysis indicated characteristic cupshaped membrane vesicles with a diameter of 50–150 nm (Fig. 2b). Western blot results confirmed exosomal markers CD63 and TSG101 detected in Exo samples, whereas calnexin and HSC70 were not detectable in Exo samples (Fig. 2c). Expression of Cav 1.2 was downregulated in CMs treated with MFBderived Exos compared with that in CMs treated with FBderived Exos (Fig. 2d, e). Collectively, these results showed that Exo, rather than the other biologically active substances secreted by MFBs, was the major component triggering repression of Cav 1.2 expression in atrial CMs.
Exos Derived from MFBs Block Adrenergic StimulationElicited Ca2+ Influx Signals in Atrial CMs
Reduced expression of Cav 1.2 in CMs suggested decreased Ca2+ influx via Cav 1.2. To test this hypothesis, we examined Ca2+ influx induced by (α,)βadrenergic agonists (Fig. 3). Since the Ca2+ signaling in CMs is thought to be mediated by Ca2+ influx via the ICaL which subsequently triggers Ca2+ release from SR, pretreatment with the SR Ca2+ATPase inhibitor thapsigargin allowed us to detect Ca2+ influx signals without contamination by Ca2+ release from SR. These results showed that either ISO or NE induced elevation in [Ca2+]i in CMs pretreated with FBderived Exos, whereas both ISOand NEinduced rise in [Ca2+]i were significantly blunted by pretreatment with MFBderived Exos.
Exosomal miR213p is a Potential Mediator Involved in the Paracrine Interaction between MFBs and CMs
We then tried to figure out the potential mechanism(s) that may be responsible for Cav 1.2 downregulation in CMs (Fig. 4). The levels of two Cav 1.2inhibitory microRNAs (miRNAs; miR213p and miR215p) selected on the basis of the published literature or bioinformaticanalysis (TargetScan, miRWalk, and RNAhybrid) were measured by qRTPCR. The results showed that miR213p levels were upregulated in MFBs and their Exos, and in microscopic analysis. n=3–4. Scale bars, 200 nm (upper panel) or 100 nm (lower panel). d Atrial CMs were treated with PBS (control) or Exos (10 μg/mL) for 24 h. Expression of Cav 1.2 protein was determined by Western blot. Data are presented as mean ± SEM. n=6. * p<0.05 versus the other two groups. e Immunofluorescence images of Cav 1.2 (red) in Hip biomechanics atrial CMs. n=3. Scale bars, 50 μm. CM, cardiomyocyte; Exo, exosome; FB,fibroblast; HSC70, constitutive heat shock protein; MFB, myofibroblast; TSG101, tumor susceptibility gene 101 protein.
Discussion
In the current study, wereport an Exomediated crosstalk between MFBs and CMs, contributing to enhanced susceptibility to AF. This work opens the way for future research into Exorelated regulatory mechanism involved in arrhythmia and may inform the development of novel predictors and therapies for AF.
Atrial fibrosis plays an important role in AF and is a final common pathway for a variety of cardiac conditions associated with AF [1]. A critical event in the development of myocardial fibrosis is the activation of FBs into MFBs [12]. Here, to establish an in vitro model that mimics tissuelevel aspects of atrial fibrosis, we used Ang II to induce FB activation to contractile MFBs characterized by de novo expression of αSMA and increased expression of extracellular matrix proteins (fibronectin and collagen) and matricellular proteins (CTGF/CCN2 and osteopontin) [20–22]. It is noteworthy that the nonAng IItreated FBs also expressed some microfilament proteins, which might be attributed to stiff culture substrates and suggested that these FBs could be referred to as protoMFBs [23]. MFBs are not present in healthy myocardium and are distinct from FBs in many respects [11, 12]. During pathological conditions, MFBs regulate ion channel expression in CMs by releasing paracrine factors, such as transforming growth factorβ1 [13–15, 24]. However, the identification of the paracrine factors involved in regulation of Cav 1.2 expression remains largely unclear.
Recently, the roles of Exo in crosstalk between cells have gained increasing attention. Previous studies have shown that Exos derived from (myo)fibroblasts trigger CM hypertrophy [16, 17]. However, whether Exomediated crosstalk between MFBs and CMs affects Cav 1.2 expression remains unknown. We found that cocultures of MFBs and CMs resulted in reduced protein expression of Cav 1.2 in CMs, which was substantially abolished by pretreatment of MFBs with exosomal inhibitor GW4869. Furthermore, treatment with MFBderived Exos decreased Cav 1.2 protein expression in CMs. Our work is the first to reveal that Exo, rather than the other biologically active substances secreted by MFBs, is the major component triggering downregulation of Cav 1.2 expression in atrial CMs.
The decreases in ICaL density and protein content of Cav 1.2 are hallmarks of AFassociated electrophysiological remodeling and major determinants of the shortening of the refractory period [4–6, 25]. Although the reduction in Cav 1.2 expression has been traditionally attributed to high beating rate during AF [7, 26], growing evidence shows that this alteration also precedes AF. The expression of Cav 1.2 is reduced inpatients [8, 9] or animal models [27, 28] with conditions that predispose to AF. However, the identification of the pathogenic factors involved in this alteration is not yet fully understood. Exomediated communication between MFBs and CMs provides a susceptible substrate that promotes AF, which implies that new therapies targeting this paracrine interaction maybe beneficial for the prevention of AF. Additionally, the presence of Exos in human pericardial fluid has been recently described, thus representing new potential clinical biomarkers [29]. Prediction of AF initiation and recurrence is highly desirable. Future work may help determine the potential predictive value of Exos derived from MFBs.
A previous study demonstrated that coculture of CMs with MFBs isolated from pressureoverload hearts leads to a decrease in Ca2+ transient amplitude in CMs [15]. Nevertheless, the effect of Exos derived from MFBs on Ca2+ response to adrenergic stimulation remains unclear. After depletion of the SR Ca2+ store by thapsigargin [18], the rise in [Ca2+]i was assumed to be due to Ca2+ influx via voltagegated calcium channels (the predominant Cav channel form expressed in CMs is the Cav 1.2 channel [19]). It is well known that both ISO and NE activate ICaL and potentiate the Ca2+ influx signals. Interestingly, both ISOand NEelicited Ca2+ influx signals in CMs were markedly attenuated by pretreatment with MFBderived Exos, which was consistent with the paralleled change in Cav 1.2 expression, and had greater arrhythmogenic potential.
It is not surprising that Exo turned out to be the major paracrine factor responsible for reduced Cav 1.2 expression. Exos have been widely thought to be quite stable and can contain proteins and miRNAs. Considering the fact that 98% of the proteins in MFBderived Exos are identical with those in FBderived Exos [17], we speculated that other molecules in Exo mediate its regulatory functions.
A variety of studies have suggested that miRNA is implicated in AF. Both miR215p and miR213p were found to be upregulated in the atrium of patients with AF [30], and computational analyses predicted that Cacna1c could be potential target for miR213p. Nevertheless, miR215p is predominantly expressed in cardiac FBs rather than in CMs [31]. Thus, to determine whether the levels of miR215p and miR213p are altered in CMs in response to Exo is of particular interest. We found that miR213p, but not miR215p, was upregulated in CMs in response to MFBderived Exos. In line, the levels of miR213p were elevated in MFBs and their Exos, indicating that exosomal miR213p maybe a mediator involved in the crosstalk between MFBs and CMs. Mechanistically, miRNAs generally bind to the complementary sequence of their targeted mRNAs and lead to their repression or degradation.
Indeed, Cacna1c was predicted to have three miRNA recognition sites in the 3′untranslated region for miR213p. Intriguingly, computational prediction also identified Akinase anchoring protein (AKAP) and protein kinase C (PKC) as potential targets. Acting as a scaffold protein, AKAPs anchor PKC or protein kinase A to Cav 1.2, thereby regulating adrenergic receptor stimulated Cav 1.2 [32]. This is consistent with our observation that adrenergicagonistinduced Ca2+ influx was abolished by MFBderived Exos. Thus, it is possible that exosomal miR213pmodulates Cav 1.2 (phosphorylation and internalization) via regulation of AKAP/PKC signaling. Further studies validating the interaction between miR213p and Cav 1.2 will lead to a better understanding of the potential regulatory mechanisms. Additionally, we also found that the level of miR215p was increased in Ang IItreated MFBs compared with that in untreated FBs. miR215p has been suggested to be important in the development of cardiac fibrotic remodeling [33]. Therefore, the present study further supports the concept that structuraland electrical remodeling processes are not independent and share some common pathways.
In summary, our work uncovers a previously unrecognized Exorelated regulatory mechanism involved in the development of a susceptible substrate for AF. Given that atrial fibrosis is irreversible, novel therapeutic approaches based on Exomediated crosstalk between MFBs and CMs may provide promising alternatives to the current treatment of AF.
