Atrial fibrillation (AF) is a common, recalcitrant-to-treatment arrhythmia associated with severe complications including thromboembolic events, heart failure progression, reduced quality of life and increased mortality (Wolf et al., 1998; Nattel and Harada, 2014; Lippi et al., 2020). The prevalence of AF doubles every decade of life and is associated with multiple comorbidities including, but not limited to, structural heart disease, arterial hypertension, obesity, diabetes and sleep apnea (Ehrlich et al., 2002; Heijman et al., 2016; Anter et al., 2017; Hayashi et al., 2017; Nattel and Dobrev, 2017). Current approaches for rhythm control in AF patients are far from optimal. Ablation therapies are invasive and limited by cost, complexity, potential life-threatening complications and uncertain long-term outcome (Ramirez et al., 2020). Available drugs are only modestly effective, and some have unfavorable side effects including predisposition to life-threatening arrhythmias (Dobrev and Nattel, 2010).

AF has a self-perpetuating and self-exacerbating nature; thus, AF patients are progressively more prone to recurrence and persistence of the arrhythmia. The susceptibility of atrial tissue to AF (i.e. AF substrate) is developed through structural and electrical remodeling in ways that are still being elucidated (Li nd Brundel, 2020; Nattel et al., 2020). Increasing evidence indicates that cardiac dysmetabolism plays a central role in the pathophysiology of AF-related atrial remodelling as well as in the AF substrate of diabetic/pre-diabetic patients (Lee et al., 2020). In fact, dysfunction of adenosine monophosphate-activated protein kinase (AMPK), which is a major regulator of cardiac metabolism, can lead to cellular abnormalities and increase AF substrate (Harada et al., 2012; Qiu et al., 2016a).

During cardiovascular distress such as ischemia-reperfusion injury, the endocannabinoid system (ECS) is activated and reportedly produces beneficial effects (Wagner et al., 2001; Montecucco and Di Marzo, 2012). The ECS is composed of cannabinoid receptors (CBR), endocannabinoid (eCB) ligands, and enzymes to biosynthesize, degrade and transport eCBs (Moreno et al., 2019). All components of the ECS are present in the heart. (Galiegue et al., 1995; Schmid et al., 2000; Bonz et al., 2003). Furthermore, in response to endogenous and external stimuli, eCBs are produced and interact with CBRs to exert regulatory actions toward the maintenance of internal homeostasis through adaptive cellular modification (Moreno et al., 2019). Therefore, exogenous cannabinoid interactions with the cardiovascular ECS and CBRs may elicit beneficial effects to modulate complications of cardiomyopathies, such as arrythmias (Montecucco and Di Marzo, 2012).

CBRs are G protein-coupled receptors of which there are two types. CBR type 1 (CB1R) predominates in the central nervous system, and is the key signaling effector of any CB-dependent psychoactive actions (Matsuda et al., 1990; Pertwee et al., 2010). In contrast, ligand activation of CBR type 2 (CB2R) typically leads to immunomodulation such as immune cell migration and cytokine release (Munro et al., 1993). Agonists that bind to CB2R lack the psychoactive effects elicited by CB1R agonists in the brain (Munro et al., 1993; Pertwee et al., 2010). CB1R and CB2R are also found in the periphery and regulate processes in organs such as the gastrointestinal tract, lungs, skin, bone, heart and liver (Gebremedhin et al., 1999; Liu et al., 2000; Bonz et al., 2003; Bouchard et al., 2003).

Exogenous cannabinoids, whether plant-derived or synthetic, have a wide range of physiological effects through interaction with the ECS. While eCB actions are presumably limited to the cellular site in which they are synthesized and released, the delivery of exogenous cannabinoids into the body results in excess compound concentrations that are more widely distributed (Moreno et al., 2019). Thus, exogenous cannabinoid effects are much more prolonged compared to eCBs (Moreno et al., 2019). Synthetic cannabinoids, originally developed to act as pharmacological probes of the ECS, are derivatives of phytocannabinoids and eCBs (Le Boisselier et al., 2017). Hundreds of varieties of new and novel synthetic cannabinoids exist, as minor modifications to the chemical structure alter the affinity and selectivity to CBRs (Le Boisselier et al., 2017).

Using CB13, a CB1R and CB2R dual agonist that does not cross the blood-brain barrier, we previously reported antihypertrophic effects and attenuated mitochondrial dysfunction in cardiomyocytes via activation of AMPK signaling pathways (Lu et al., 2014; Lu et al., 2020). We hypothesize that the beneficial effects of CB13 may be relevant to stressed atria as well. Here, we used a Langendorff-perfused rat heart to assess the effects of CB13 on basic physiology of the heart as well as on the atrial electrophysiology following acute atrial tachypacing to mimic an atrial tachyarrhythmia. Our findings support the notion that the beneficial effects of CB13 are indeed relevant in the stressed atria and may serve as new therapeutic strategy to attenuate AF-related atrial remodeling.

The present study extends extant evidence indicating possible beneficial effects of ECS activation in the stressed myocardium (Wagner et al., 2001; Lu et al., 2014; Lu et al., 2020). Overall, this study demonstrated three salient findings regarding the cardiac effects of the peripherally restricted, dual CBR agonist CB13 in rats: 1) lack of CB13-dependent chronotropic, dromotropic or hemodynamic effects in the non-paced ex-vivo preparation; 2) an ability of CB13 to antagonize the AERP remodelling induced by acute atrial tachypacing; 3) AMPK activation may contribute to the ability of CB13 to protect against tachypacing-induced atrial remodelling. These topics are further discussed in detail below.

Our findings address the paucity of knowledge regarding the direct effects of CBR agonists on the intrinsic properties of the intact heart (Weresa et al., 2019), and support the notion that combined activation of CB1R and CB2R probably does not markedly alter the basic electrophysiology, calcium handling and electromechanical coupling of the normal rat myocardium. Indeed, our data indicate that CB13 treatment does not alter chronotropy, dromotropy, or hemodynamics in the spontaneously beating rat heart. Activation of solely CB1Rs reportedly induces negative inotropic effects in isolated human atrial muscle (Bonz et al., 2003). Likewise, the CB1R agonist, HU-210, induced negative inotropic effects in the LV of the isolated rat heart (Krylatov et al., 2005), although a recent study also using HU-210 did not report such effect (Gorbunov et al., 2016). Interestingly, in the presence of the CB1R antagonist AM251, anandamide exerted a CB2R-mediated positive inotropic effect in rat atria (Sterin-Borda et al., 2005). Thus, it is conceivable that the absence of CB13 effects on intrinsic cardiac properties results from opposing effects of CB1R activation and CB2R activation which manifest as a net zero effect. This is consistent with previous reports of opposing CB1 vs. CB2 effects. In fact, CB2R activation has been linked to cardioprotective effects including anti-arrhythmic effects (Krylatov et al., 2001), whereas CB1R activation leads to a myriad of cardiodeleterious effects (Mukhopadhyay et al., 2007; Dol-Gleizes et al., 2009; Sugamura et al., 2009; Mukhopadhyay et al., 2010; Rajesh et al., 2010; Tiyerili et al., 2010; Rajesh et al., 2012; Molica et al., 2013). Thus, Lu et al proposed that, in the context of attenuating cardiac myocyte hypertrophy, activation of CB2Rs is necessary to negate any adverse effects produced by CB1Rs. Further pharmacological studies will be required to systematically explore this possibility in terms of chronotropy, dromotropy, or hemodynamics in the spontaneously beating rat heart.

Atrial tachypacing is an important experimental approach to mimic AF-related remodeling. In large mammals, sustained tachypacing modifies atrial properties so that AF susceptibility and sustainability gradually increase over time (Wijffels et al., 1995; Nattel et al., 2020). Studies have indicated that the rapid atrial activity and consequent cellular calcium overload within atrial cardiomyocytes lead to secondary changes which converge to increase AF substrate. Shortening of AERP is a hallmark of AF-related remodeling and an important factor that promotes multiple circuit reentry in the atria (Wakili et al., 2011; Nattel et al., 2020). In rodents the long-term effects of atrial tachypacing are less clear. However, short-term tachypacing consistently leads to AERP shortening (Etzion et al., 2008). In addition, we recently demonstrated that atrial tachypacing of freely moving rats increases AF substrate and promotes molecular changes that resemble those reported in large mammalian models (Mulla et al., 2019). Moreover, tachypacing of atrial cells in vitro leads to molecular changes that recapitulate those observed in the atrial tissue of AF patients (Wiersma et al., 2017; Zhang et al., 2019). Our current findings indicate that CB13 inhibits the AERP shortening induced by acute atrial tachypacing in the ex vivo rat heart. This finding supports the notion that activation of CBRs may be a new therapeutic target in the context of AF-related remodelling. However, further studies are needed to delineate the long-term effects of this new therapeutic strategy in vivo. In addition, since AF susceptibility of the normal rat heart is very low, it is challenging to induce AF even following atrial tachypacing and AERP shortening in the current experimental setup. Therefore, the effect of CB13 on AF induction will rely on future preparations that are more prone to AF induction (e.g., the atria of rats with heart failure post myocardial infarction) (Klapper-Goldstein et al., 2020).

The biochemical findings of our study suggest that AMPK activation is involved in the protective effect of CB13 against tachypacing-induced atrial remodeling. AMPK is an essential signaling molecule that contributes to the maintenance of intracellular ATP levels (Dolinsky and Dyck, 2006). To this end, AMPK controls energy homeostasis and senses changes in AMP:ATP to in turn control ATP-consumption and catabolic pathways(Dolinsky and Dyck, 2006). Reduced AMPK activation in atrial cardiomyocytes appears to be deleterious in regards to regulating metabolic and oxidative stress (Ozcan et al., 2015). Importantly, multiple risk factors for AF such as heart failure, myocardial ischemia, and hypertrophy are associated with metabolic stress and cellular dysfunction (Harada et al., 2012). By virtue of ameliorating these metabolic abnormalities, AMPK activation may thus be implicated in improving arrhythmias (Harada et al., 2012). In fact, Ozcan et al demonstrated that liver kinase-B1 (LKB1; an AMPK kinase that is linked to activation of AMPK) knockout mice exhibit impaired contraction and decreased AMPK activity (Ozcan et al., 2015). This LKB1 knockout model replicates human disease progression as it develops spontaneous AF that develops into persistent AF (Ozcan et al., 2015).

Our data indicate that acute atrial tachypacing ex vivo leads to downregulation of phosphorylated AMPK, similar to that found in humans experiencing persistent AF (Harada et al., 2015). In this context, CB13 abrogated AMPK downregulation and increased phosphorylation at its activation site, presumably improving metabolic and cellular function. Although we did not demonstrate a direct link between the electrophysiological effects and metabolic findings, we speculate that improved metabolic conditions of atrial cells in the presence of CB13 lead to less oxidative stress-dependent AERP shortening (Korantzopoulos et al., 2007; Yoshizawa et al., 2018). However, this hypothesis will have to be addressed more directly in future studies. Furthermore, AERP reduction may also be caused by other factors; for example Ozgen et al demonstrated involvement of Ca²⁺-activated K⁺ currents in action potential duration and atrial remodeling (Ozgen et al., 2007). Additionally, AMPK activation by CB13 is likely not mediated through LKB1 activation (Figure 5B) and is therefore likely achieved through changes in cellular Ca²⁺ levels, particularly since Ca²⁺/calmodulin-dependent protein kinase kinases (CAMKKs) and reactive oxygen species (ROS), which promote AF, have been implicated to alter AMPK signalling (Lin et al., 2004; Kim et al., 2016). In fact, CAMKKα and CAMKKβ phosphorylate AMPK at its activation site (Woods et al., 2003; Hawley et al., 2005; Hurley et al., 2005), and CBR ligands can signal through CAMKKs (Vara et al., 2011). An important future direction will be to utilize a CAMKK inhibitor to validate this hypothesis.

Information regarding the mediator between AMPK and arrhythmia development, tissue remodeling, or modification of AF substrate remains limited(Harada et al., 2012). However, there does appear to be a relationship between cardiac EP and metabolic changes that involve a key role of AMPK and AF substrate (Qiu et al., 2016a; Qiu et al., 2016b). In the current experimental setup, we identified PGC-1α as a candidate mediator for CB13-dependent effects (Figure 5C). PGC-1α regulates mitochondrial function vis-à-vis its function as a key transcriptional co-activator that modulates mitochondrial biogenesis, ATP synthesis, and reactive oxygen species (ROS) defense mechanisms (Ventura-Clapier et al., 2008; St-Pierre et al., 2006). Activation of AMPK results in a downstream cascade that increases PGC-1α expression (Jager et al., 2007). Notably, PGC-1α overexpression rescues cardiac mitochondrial function (Zhou et al., 2013). Tachypacing had no effect on PGC-1α compared to non-pacing controls (Figure 5C). However, CB13 treatment caused a significant upregulation of PGC-1α. We previously speculated that CB13 activates AMPK via CB2R whereas CB1 receptors invoke other signaling pathways (Lu et al., 2020). In addition, Zheng et al. demonstrated that JWH-133, a CB2R-selective agonist, adequately activated AMPK to stimulate PGC-1α, without CB1R interaction (Zheng et al., 2013). Here we elucidated that CB13 treatment during tachypacing induces a CBR/AMPK (possibly CB2R, specifically) signaling cascade, resulting in PGC-1α activation. While we have previously shown that CB13 has altered mitochondrial bioenergetics and mitochondrial membrane potential in hypertrophied ventricular cardiomyocytes, an important next step is to determine the effects of CB13 on mitochondrial dysfunction in atrial cardiomyocytes, and the role of the AMPK signaling cascade therein.

Lastly, our data indicate that tachypacing reduces the atrial expression of Cx43, and this reduction was attenuated by with CB13 treatment. Cx43 is a major gap junction in the connexin family and is abundantly found within the atrial and ventricular myocardium (Severs et al., 2008). Cx43 has been investigated in several AF-related remodelling studies; for example, in dog pacing models Cx43 reduction has been demonstrated after tachypacing for 3 days, with reduction being preserved for up to 21 days (Akar et al., 2007). In rabbit atrial tachypacing experiments Cx43 reduction has been shown to promote atrial fibrillation development (Yan et al., 2013). Additionally, Ozcan et al demonstrated that AMPK activation may inhibit atrial fibrillation in LKB1 knockout mice, and similarly found LKB1 knockout mice had a reduction in Cx43 that was rescued by AMPK activation (Ozcan et al., 2020). In contrast, Alesutan et al demonstrated inhibition of Cx43 by AMPK (Alesutan et al., 2015); thus Cx43 and its interaction with AMPK in the context of electrical remodeling remains inconclusive. Furthermore, downregulation of Cx43 would lead to reduced EP coupling between atrial myocytes, which in turn would slow conduction velocity and increase AF susceptibility in atrial tissue (Luo et al., 2018). While we did not demonstrate a change in conduction velocity between groups there was a trend towards conduction lengthening in the vehicle treatment group. It is possible that the short-term tachypacing in our model did not allow for more prominent effects to be observed. Of note, oxidative stress associated with rapid atrial activity has been shown to cause Cx43 alterations (Boengler and Schulz, 2017; Nattel et al., 2020); and this may be modulated by AMPK activation (Qiu et al., 2016b). Further studies to examine the interplay between CB13, AMPK, and Cx43 are clearly warranted.

In conclusion, our data demonstrating the ability of CB13 treatment to prevent tachypacing-induced AERP shortening and Cx43 reduction, as well as to activate known cardioprotective signaling mediators in terms of metabolic dysfunction (i.e., AMPK and PGC-1 α), are collectively in favor of a putative beneficial effect of this drug as a new upstream modality that might be utilized to prevent atrial electrical remodeling. Considering the possible beneficial effects of such treatment in the context of pathological hypertrophy (Lu et al., 2014; Lu et al., 2020), which shares many of the risk factors for AF, this therapeutic strategy is even more attractive for further investigation.