Arrhythmogenic Right Ventricular Cardiomyopathy (ARVC) is an inheritable condition characterized by a fibrofatty infiltration of the heart muscle (predominantly, but not exclusively, the right ventricular free wall), as well as increased risk of ventricular arrhythmias and sudden cardiac death (SCD)(Goff & Calkins, 2019; Jordà et al., 2022). Most cases of gene-positive ARVC associate with pathogenic variants in the gene PKP2, coding for the desmosomal protein Plakophilin-2 (PKP2)(Bhonsale et al., 2015). The disease phenotype progresses over time, with lethal arrhythmias often observed in the early stages, before the onset of the cardiomyopathy(Groeneweg et al., 2014). While studies regarding the arrhythmogenic substrate in ARVC have focused on the risk of life-threatening ventricular arrhythmias, recent studies have drawn attention to the high prevalence of atrial fibrillation (AF) among ARVC patients. Recent studies show that that the incidence of atrial arrhythmias ranges between 15–25% in ARVC, and that AF is the most prevalent supraventricular arrhythmia (59%) among patients(Camm et al., 2013; Cardona-Guarache et al., 2019; Baturova et al., 2020). Multiple biobank-driven genome-wide association studies recently revealed an association of AF with genetic variants in PKP2, further supporting the notion that PKP2 is important for atrial function(Nielsen et al., 2018; Zöller et al., 2024; Roberts et al., 2024). Atrial arrhythmias are clinically important for ARVC patients, as they associate with inappropriate implantable cardioverter-defibrillator shocks, increased risk of sustained ventricular arrhythmias and disease morbidity(Camm et al., 2013; Cardona-Guarache et al., 2019). However, the pathophysiology behind the development of atrial arrhythmias in ARVC and specifically, the association between PKP2 deficiency and atrial arrhythmogenic mechanisms, remains understudied.

It is generally accepted that both a substrate and a trigger are essential for the initiation and maintenance of arrhythmic events(Cluitmans et al., 2023). For ventricular arrhythmias in PKP2-deficient hearts, dysregulation of intracellular calcium (Ca²⁺_i) homeostasis appears to be a critical trigger(van Opbergen et al., 2022), while a disrupted micro-structure may be sufficient substrate for arrhythmic events(Corrado et al., 2023). In the present study, we examined potential triggers and substrate elements that could affect the stability of the atrial rhythm in hearts deficient in PKP2. Our studies utilized a well-characterized model of a cardiac-specific, tamoxifen (TAM)-activated knockout of PKP2 (PKP2cKO) as well as an analysis of a database for genetic associations in humans (GTEx). Our experiments followed the path previously traced by our studies on PKP2-related ventricular dysfunction(Cerrone et al., 2017; Pérez-Hernández et al., 2022). Features such as fibrosis, excess ROS production and Ca²⁺_i dysregulation were also observed in atrial myocardium, though important differences became apparent. An in-depth transcriptomics analysis revealed novel elements in the molecular framework supporting the atrial phenotype. Our observations suggest that recent therapeutic approaches to mitigate ventricular ARVC(van Opbergen et al., 2022; Wojciech Zareba et al may be effective to treat the atrial disease.

The data that support the findings of this study are presented in the article and online supplementary files. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE(Perez-Riverol et al., 2022) partner repository with the dataset identifier PXD052074. Additional information can be made available from the corresponding author upon reasonable request.

Reviewers can access the proteomics data with the following credentials:

Username: reviewer_pxd052074@ebi.ac.uk Password: 8Lo9drdN

We examined the electrophysiological profile, and the basic parameters reflecting Ca²⁺_i homeostasis, in PKP2cKO atrial myocytes harvested 21 days post injection (21 dpi) of TAM, a time point at which an arrhythmogenic cardiomyopathy of right ventricular predominance can be observed(Cerrone et al., 2017). As shown by the data in Figure 1 we detected no significant changes in the morphology of the action potential, nor did we identify an effect on peak density or steady-state inactivation of the sodium current. Moreover, and in sharp contrast to what we have observed in ventricular myocytes harvested at the same time point after TAM(Cerrone et al., 2017), we detected no differences in the amplitude of elicited Ca²⁺_i transients or in the time course of Ca²⁺_i decay after activation when comparing control to PKP2cKO cells (Figure 2A&B; bars labelled “control” and “PKP2cKO”). In addition, acute exposure to isoproterenol (ISO) elicited an increase in the amplitude of the transient and a reduced time constant of transient decay of similar magnitude in both cell populations (Figure 2A&B; bars labelled “Control + ISO and PKP2cKO + ISO”). Yet, we observed a large increase in the frequency and amplitude of Ca²⁺ sparks in PKP2-deficient cells (comparing “control” versus “PKP2cKO”). PKP2cKO cells were also more susceptible to modulation by beta-adrenergic stimulation (Figure 3A&B; other parameters in Fig 4A&B; comparison effect of ISO in Figure 5A&B). This pattern was consistent with the observed measurements of diastolic Ca²⁺ levels, presenting a significant increase consequent to ISO stimulation in the PKP2cKO cells (Figure 6A; comparison effect of ISO in Figure 5C). SR load increased by ISO exposure in both control and PKP2cKO cells; the magnitude of the SR load after ISO was significantly larger in control than in PKP2cKO cells, likely because of excessive leak of Ca²⁺ out of the SR in the latter group (Figure 5DE & 6B).

The results above suggest an “eagerness” for the RyR2 channels in PKP2cKO ACMs to flicker and release Ca²⁺ into the cytoplasmic space during diastole. Further evidence was obtained through acute in vivo treatment of PKP2cKO mice with Dantrolene, a well-characterized ryanodine receptor blocker(Kobayashi et al., 2009). Data are shown in Figure 3&6. PKP2cKO mice were injected with either DMSO or Dantrolene (30mg/kg) 30 minutes prior to euthanasia (see also Methods). The results show that Dantrolene mitigated the ISO-dependent increased frequency of Ca²⁺ sparks and their amplitude in PKP2cKO cells (Figure 3A–C). A corresponding mitigation of diastolic Ca²⁺ elevation and increase in SR load was observed (Figure 6A&B).

Previous studies have indicated that the open probability of RyR2 channels can be modulated by its phosphorylation state, particularly in Ser-2808 and Ser-2814(Wehrens et al., 2006; Uchinoumi et al., 2016). We therefore utilized mass spectrometry analysis to evaluate for abundance differences of RyR2 phosphorylation sites in control and PKP2cKO atria. We obtained excellent coverage of Ryr2 phosphorylation sites; our data revealed no difference between groups (Figure 7A), thus indicating that the increased frequency and amplitude of the Ca²⁺ sparks could not be accounted for by changes in RyR2 phosphorylation in those sites. On the other hand, reactive oxygen species (ROS), known to post-translationally modify RyR2 oxidation state and increase open probability, were more abundant in PKP2cKO myocytes when compared to control (Figure 7B–C).

Loss of PKP2 leads to enhanced sarcomere shortening in adult ventricular myocytes(Cerrone et al., 2022). Those results were interpreted as consequent to the increased amplitude and duration of Ca²⁺_I transients. Interestingly, a similar analysis of sarcomere shortening in atrial PKP2cKO myocytes, which show no difference in Ca²⁺_i transient amplitude or duration, also revealed that loss of PKP2 expression associates with increased shortening, reduced sarcomere length at rest and a prolonged relaxation time when compared to control (Figure 8A–E). Yet, while isolated myocytes were able to maintain contractility, we observed a drastic loss of muscle integrity at the expense of a large increase in collagen abundance in the PKP2cKO atria when compared to control (Figure 9A&B). These results were similar to those previously reported for ventricular tissue, where loss of PKP2 expression associated with a loss of muscle mass (and in the case of the left ventricle, reduced ejection fraction) while the fractional shortening of individual myocytes was in fact, enhanced(Cerrone et al., 2022). Finally, as suspected, the structural remodeling in the macro- scale was also apparent at the micro-scale, as PKP2cKO in atrial myocytes led to the expected loss of intercalated disc integrity, as shown by the electron microscopy images presented in Figure 9C and in Supplemental Movies 1&2.

To further characterize the effects that loss of PKP2 exert on the molecular profile of the atria, we conducted a differential transcriptome analysis comparing 6 atria from PKP2cKO hearts versus 6 atria from controls. The differential transcriptome is presented in the form of a volcano plot in Figure 10A. Positive and negative numbers in the x-axis reflect the magnitude of the differential, expressed as Log2Fc. Only 215 transcripts were significantly down-regulated and 136 were up-regulated in PKP2cKO samples (| log2FC | >0.25, FDR < 0.05). As expected, given the nature of the experimental model, the largest change was found in the transcript abundance of Pkp2 (downregulated) and the estrogen receptor Esr1 (upregulated). The complete dataset is available as Supplemental Table 1. As a next step, we sought to discover functionally relevant co-regulated transcripts in the human transcriptome. For this purpose, we took advantage of the GTEx repository.

We implemented a previously described protocol(Montnach et al., 2018) to explore the relation between the abundance of the Pkp2 transcript and the abundance of each of the 27,856 transcripts stored in the GTEX v8 databank under the human atria category. The pipeline for data generation is shown in Figure 11. The resulting volcano plot (Figure 10B) presents two arms: one, to the right of the ordinate, indicating that the abundance of Pkp2 correlated directly with the abundance of a given transcript (a positive regression coefficient in the linear relation; see Figure 11) and the one to the left of the ordinate, where the abundance of Pkp2 is inversely correlated with the specific transcript (negative regression coefficient). Overall, the plot shows a high degree of association between the abundance of Pkp2 and that of 12,611 genes with a slope of the linear regression fit of more than 0.5 and an adjusted p-value (padj) value of <0.001, of which 5,419 transcripts were positively correlated and 7,192 transcripts were negatively correlated. Pathway analysis of those genes is shown in Supplemental Table 2. We found consistency with pathways previously identified from bulk transcriptomics of PKP2cKO tissue(Cerrone et al., 2017) and of ARVC human samples(Pérez-Hernández et al., 2022) and that have been linked to the pathophysiology of ARVC, such as oxidative phosphorylation, citrate cycle (TCA cycle), and nucleotide excision repair (Figure 12). We also found that, consistent with previous studies, there is an over-representation of pathways linked to neurodegenerative diseases (Figure 12B).

The nature of the GTEx plot is only correlative. The differential transcriptomic data (via RNA-seq of PKP2cKO vs control atria) on the other hand, directly reflects changes in the abundance of a given gene consequent to reduced PKP2 expression. We therefore cross-referenced both datasets to identify transcripts that show a significant correlation with PKP2 abundance in the human heart and that depend on PKP2 expression, as determined by the RNA-seq experiments. We found that 33 of the 215 downregulated transcripts in the PKP2cKO mouse had a significantly positive correlation with Pkp2 (blue circles in Figure 10B) and 8 of the 136 upregulated transcripts had a significantly negative correlation with Pkp2 (red circles in Figure 10B). The complete datasets are provided in Supplemental Table 3. Our results suggest that genes involved in the pro-fibrotic remodeling of the heart (e.g. COL4A4, COL8A1, COL18A1 and P3H3), in an intercalated disc-transcription axis (JUP) and in mitochondrial function (e.g. NDUFB4, COX7A2, ATP5MD, DNAJC15 and NDUFB1) are part of a common gene network that is subject to modulation by the levels of Pkp2 expression. Interestingly, this network also includes genes coding for proteins that regulate sarcomere function (e.g. MYBPC3, ITGA7, SGCA and LRRC39). As a final step, we tested whether the prediction obtained through the discovery process outlined in Figure 10, was confirmed by direct experimental methods. We focused on MYBPC3, a transcript coding for a protein critical for the regulation of sarcomere function in the adult heart and related to several cardiomyopathies, as well as atrial fibrillation((Hershberger et al., 2010; Owais et al., 2024)). As shown in Figure 13, RT-PCR (A), Western blot (B) and immunolocalization (C) experiments revealed a reduced abundance of the MyBPC3 protein in PKP2cKO atrial myocardium and isolated PKP2cKO atrial myocytes when compared to control. These results were consistent with those obtained through the transcriptomic pipeline and supported the notion of a link between PKP2 and MyBPC3 that can have important consequences in the generation of the atrial cardiomyopathy observed in PKP2 deficient hearts.

We carried out a comprehensive characterization of the role of PKP2 in the atrial myocardium, using a model of PKP2 deficiency (PKP2cKO mice) with a well-characterized ventricular phenotype. In contrast to the previously-reported observations in PKP2cKO ventricular myocytes(Cerrone et al., 2017), atrial PKP2cKO myocytes presented no significant differences in Ca²⁺ transient dynamics, SR load and action potential morphology. However, PKP2cKO atrial myocytes showed an increased frequency and amplitude of Ca²⁺ sparks, in combination with increased diastolic intracellular Ca²⁺ and cellular ROS levels. ISO exposure further increased the frequency of Ca²⁺ sparks and the levels of diastolic Ca²⁺, and those effects were reversed by acute in vivo treatment with Dantrolene, a known RyR blocker. PKP2cKO atrial myocytes also presented enhanced sarcomere shortening and impaired sarcomere relaxation. Transcriptomic analysis followed by direct biochemical studies identified a reduced abundance of the sarcomeric protein MyBPC. The latter, we speculate, may have a mechanistic role in the cellular shortening and relaxation changes observed. A comparison between the changes in PKP2cKO ventricular versus atrial myocytes is presented in Figure 14.

Fibrotic infiltration of the ventricular wall is well-documented in PKP2cKO mice. As reported by Cerrone et al(Cerrone et al., 2017), a significant increase in collagen content is first noted in the RV free wall 21 days after TAM injection (21 dpi), eventually being also manifest in the LV free wall (at 28 dpi; see(Cerrone et al., 2017)). For the present study, mice were euthanized and studied at 21 days post-TAM. This time point has been chosen in previous reports from our lab ((Cerrone et al., 2017; van Opbergen et al., 2022, 2024; Pérez-Hernández et al., 2022)) because of a consistent characteristic: RV dilation, RV fibrosis, and yet, minor LV fibrosis with preservation of ejection fraction; in other words, a cardiomyopathy of right ventricular predominance with only minor involvement of the left ventricle. The data presented here show abundant collagen infiltrates in the left atria at 21 dpi, thereby preceding LV fibrosis and also preceding a reduction of left ventricular ejection fraction below 45%(Cerrone et al., 2017). Left atrial fibrosis seems to appear even at an earlier stage, prior to significant increase in collagen abundance in the ventricular walls, as hinted by a close analysis of previously published results (see Figure 1 of Cerrone et al; heart at 14 dpi(Cerrone et al., 2017)). These results suggest that atrial remodeling is not consequent to a failure of LV contractile function, but rather, a process intrinsic to the loss of PKP2 in the atria itself. Whether the fibrotic process contributes, through direct or paracrine regulation, to the observed changes in Ca²⁺_I homeostasis, remains unclear. On the other hand, it is possible that myocyte-intrinsic changes in contractility, ROS production and Ca²⁺_I dysregulation (e.g, Fig 2,3,6,7&8) favor myocyte damage and myocyte death, which would be followed by a reparative fibrosis throughout the structure. Both processes in combination (a cell-intrinsic pro-fibrotic process consequent to PKP2-dependent fluctuations in the transcriptional network on one hand, and myocyte damage produced by the proliferative fibroblasts, on the other) would generate a positive feedback loop that would facilitate the damage of the atria as a whole.

Exercise negatively impacts ARVC progression and is considered an independent risk factor for arrhythmias consequent to PKP2 deficiency.(Sawant et al., 2014) In ARVC patients, sympathetic input can be particularly arrhythmogenic and sympathectomy is an effective method of treatment in some patients.(Assis et al., 2019) However, the effect of catecholaminergic input on the PKP2 deficient atrial myocardium remains understudied. Here, we show that beta-adrenergic stimulation (by ISO) of PKP2-deficient atrial myocytes elicited an enhanced increase in the frequency and amplitude of Ca²⁺ sparks, compared to control cells. PKP2cKO cells were also more susceptible to beta-adrenergic stimulation in regard to diastolic Ca²⁺ levels, presenting a larger increase consequent to ISO stimulation, while the SR load upon ISO exposure tended to be lower in PKP2cKO compared to control cells. These data suggest that adrenergic stimulation of PKP2-deficient atrial myocytes causes an excessive leak of Ca²⁺ out of the SR, a well-known pro-arrhythmic trigger in cardiac myocytes. We thereby speculate that exercise, besides increasing the susceptibility for ventricular arrhythmias, as well enhances the risk for atrial arrhythmias in patients with PKP2-related ARVC.

Our previous studies show that in PKP2 deficient ventricular tissue, dysregulation of intracellular Ca²⁺ homeostasis is based on transcriptional changes of Ca²⁺ handling related genes. Yet, a similar transcriptional remodeling was not observed in PKP2 deficient atrial myocytes. Moreover, substantial structural and functional differences exist between atrial and ventricular myocytes regarding Ca²⁺ handling. Atrial myocytes lack substantial T-tubules, and possess two populations of RyRs: one minor group (junctional RyRs) that sits just beneath the sarcolemma and another group (non-junctional RyRs) located deep inside the cell and not forming a dyadic structure with sarcolemmal calcium channels(Mackenzie et al., 2004; Bootman et al., 2006). Importantly, previous studies showed that, in PKP2cKO ventricular myocytes, there is increased proximity between Ca_V1.2 and RyR2 and this proximity can play a role in the observed increase in the Ca²⁺ transient amplitude(Cerrone et al., 2017). Thus, the structural differences between atrial and ventricular myocytes may explain, at least in part, the differences in terms of Ca²⁺ transient dynamics.

Ca²⁺ reuptake into the SR is more robust in atrial myocytes, yielding a relatively greater content of Ca²⁺ in the SR when compared to ventricular myocytes(Hove-Madsen et al., 2004; Mackenzie et al., 2004; Bootman et al., 2006; Neef et al., 2010; Voigt et al., 2012). At first glance, an increase in SR Ca²⁺ leak (as observed in our studies) would be expected to reduce the SR Ca²⁺ pool; yet, no changes in SR Ca²⁺ load were detected. This seemingly paradoxical result is reminiscent of previous results showing that the CASQ2-K180R mutation provokes spontaneous RyR2 Ca²⁺ release, without affecting SR load or the expression of Ca²⁺-handling related proteins(Wleklinski et al., 2022). In that study, CASQ2-K180R was specifically found to impair the dynamic buffering of Ca²⁺ within the SR, without affecting total SR Ca²⁺ content. It would be interesting to speculate that PKP2 deficiency as well impacts on the intra-SR Ca²⁺ buffering in atrial myocytes. In addition, it is important to highlight that mitochondria influence excitation-contraction coupling by means of Ca²⁺ uptake, buffering and release(Dedkova & Blatter, 2013). Changes in mitochondrial Ca²⁺ levels due to mitochondrial dysfunction can potentially influence diastolic Ca²⁺, RyR2 function and SR load in PKP2cKO atrial myocytes by means other than (or in addition to) elevated ROS levels. Dissecting the contribution of individual components to eh overall change in RyR2 activity will be a complex and interesting endeavor in future studies. Our data show that Dantrolene was very effective in reducing the extent of spontaneous SR Ca²⁺ release events and diastolic Ca²⁺ levels, while improving SR Ca²⁺ load, upon ISO exposure. However, Dantrolene was not as effective under control (non ISO-stimulated) conditions. When β-adrenergic agonists are applied to atrial myocytes, there is a rapid change in spatial properties of the intracellular Ca²⁺ signals, whereby Ca²⁺ signaling acutely changes from being a peripheral-only elevation of Ca²⁺ to a global response in which the non-junctional RyRs participate(Mackenzie et al., 2004). The involvement of a larger pool of RyR2 channels possibly explains the more prominent effect of Dantrolene upon ISO exposure. The beneficial effect of Dantrolene on atrial Ca²⁺ dysregulation has been confirmed in human heart failure and AF cardiomyocytes(Hartmann et al., 2017) and, combined with our data, suggest that Dantrolene could be a potential antiarrhythmic for both atrial and ventricular arrhythmias in ARVC patients carrying PKP2 mutations. Dantrolene is a clinically approved drug, currently used for the treatment of malignant hyperthermia(Krause et al., 2004), and therefore amenable for repurposing to treat the ARVC patient population. It is important to note that our studies characterize the development of potential substrates and triggers in the setting of PKP2 deficiency in atrial tissue, but the anatomical differences between mice and human preclude us from making an assessment of atrial fibrillation in mice as a relevant parameter in the context of the human disease.

Our GTEx dataset suggested that PKP2 expression is coordinated with genes of importance for cardiomyocyte sarcomere function, mitochondrial energy production and pro-fibrotic remodeling of the heart. We then complemented GTEx with transcriptomic data from PKP2cKO mice, allowing us to identify not only transcripts whose abundance seem to concur, but also that their abundance depends on the expression of PKP2. Through this process, we showed that PKP2 expression is coordinated with (a) the expression of a transcript involved in cell-cell contact and the signal transduction from the intercalated disc to the nucleus (JUP), (b) transcripts involved in sarcomere function (MYBPC3, SGCA and LRRC39), (c) oxidative phosphorylation (e.g. NDUFB4, COX7A2, ATP5MD, DNAJC15 and NDUFB1) and (d) collagen syntheses (e.g. COL4A4, COL8A1, COL18A1 and P3H3). The broad discovery (omics) approach thus pointed out to individual candidates to be tested by direct biochemical methods. Follow-up studies confirmed the reduction in MyBPC3 transcript and protein and as such, bases for the hypothesis that loss of PKP2 leads to mechanical dysfunction of the atrial myocytes, and that the latter results, at least in part, from changes in the transcript abundance of an important sarcomere protein.

Mutations in MYBPC3 can lead to various forms of cardiomyopathies, such as dilated cardiomyopathy, hypertrophic cardiomyopathy and left ventricular noncompaction cardiomyopathy(Hershberger et al., 2010; Yuan et al., 2017; Tudurachi et al., 2023). In addition, a recent genome-wide association study showed that genetic variants in MYBPC significantly relate to atrial fibrillation, confirming the importance of MYBPC for atrial function(Owais et al., 2024). The protein MyBPC associates with both actin and myosin and as such, serves as a brake that limits cross-bridge interactions during contraction. Our data shows that PKP2cKO atrial myocytes present impaired relaxation and enhanced sarcomere shortening, consistent with down-regulation of MyBPC3. It is worth mentioning that previous studies have reported seemingly contrasting findings regarding the regulatory effect of MyBPC3. On one hand, Fraysse et al. showed that cardiomyocytes of mice carrying a MYBPC3 point mutation, frequently associated with HCM, exhibited higher myofilament Ca²⁺ sensitivity, shorter diastolic sarcomere length, faster Ca²⁺ transient decay and reduced in vivo fractional shortening(Fraysse et al., 2012). On the other hand, Toepfer et al. showed that stepwise loss of MyBPC3 resulted in augmentation of myosin contractility and prolongation of the sarcomere shortening relaxation time(Toepfer et al., 2019). Of note, these studies were performed in ventricular myocytes; yet, the results seem to concur with the changes in sarcomere shortening dynamics observed in PKP2 deficient atrial cardiac myocytes.

We acknowledge that the genetic deficit in PKP2 created for the purpose of this study does not reproduce that of humans with ARVC, where PKP2 deficiency is very rarely recessive and it does not affect only one cell type or developmental stage. In fact, no experimental model, be it murine or cell-based, properly recapitulates ARVC. In this study, we identify the atrial endophenotype of a protein that, when mutated in a human heart, causes ARVC. It is important to note that existing evidence suggests that PKP2-related ARVC is consequent to a loss of function of the PKP2 protein(Dries et al., 2021). The PKP2cKO model has been used to generate pre-clinical data toward actionable treatment in humans (e.g., a pilot clinical trial on use of flecainide in ARVC patients; NCT03685149(Wojciech Zareba et al., 2023), and PKP2-related gene therapy; RP-A601(van Opbergen et al., 2024) and(Wu et al., 2024)). Future studies will allow us to discern the extent of the similarities between the murine and the human atrial phenotype, and whether the murine model is also a useful tool to advance atrial-related therapy in humans.

In conclusion, we describe that loss of PKP2 expression affects cellular Ca²⁺ handling and electrophysiology differently in the atrial versus ventricular myocardium. We speculate that an increased probability of opening of Ryr2 channels, supported by increased ROS production and hence, likely RyR2 oxidation, serve as a pro-arrhythmic trigger that, combined with a highly fibrotic substrate, can facilitate atrial arrhythmias. We also found that the dysregulation of Ca²⁺_i can be amplified by catecholaminergic input and that RyR blockers can dampen these pro-arrhythmic effects, and we describe a PKP2-related gene network that includes various components of the ARVC molecular phenotype. Taken together, our findings provide novel insights into the role of PKP2 in the atrial myocardium with potential implications to the therapeutic management of AF in patients with PKP2-related ARVC.