Induction of heat shock proteins prevents the arrhythmogenic substrate for atrial fibrillation

Abstract

Atrial fibrillation (AF) is the commonest arrhythmia. Studies have shown that atrial tachypacing (artificial persistent AF) causes electrical remodelling. This is characterised by the shortening of the atrial effective refractory period (ERP), in which reduction in L-type Ca²⁺ channel current plays an essential part. Atrial fibrosis, a feature of structural remodelling, is induced by continuous infusion of angiotensin II, and has been associated with conduction delay in atria, which promotes AF. Acute atrial ischaemia, frequently observed during development of acute coronary syndrome, has been associated with atrial conduction heterogeneity, which also promotes AF. Induction of heat shock proteins (Hsp72 and Hsp27) by hyperthermia and/or geranylgeranylacetone has demonstrated to protect the heart against such atrial remodelling. The potent protective role of Hsp72 and Hsp27 against clinical AF in patients who underwent open heart surgery has been shown. Taken together, interventions that induce heat shock responses (including induction of Hsp72 and Hsp27) may prevent newly developed AF and delay the progression of paroxysmal AF to persistent AF.

• atrial refractory period
• atrial fibrosis
• atrial fibrillation

Introduction

Atrial fibrillation (AF) is the commonest arrhythmia. AF is a strong risk factor for stroke, the most devastating complication of AF [1]. The ‘downstream’ approach targeting ion channels using antiarrhythmic drugs has been demonstrated to be limited because antiarrhythmic drugs are effective in only about 50% of patients, and are often associated with adverse effects, particularly proarrhythmia [2]. An ‘upstream’ approach targeting the processes involved in the development of the substrate that promotes AF has recently attracted much attention [3]. Activation of the renin–angiotensin system, inflammation, and oxidative stress are involved in development of the AF substrate, and are therefore upstream therapeutic targets [3]. A growing body of evidence suggests that induction of heat-shock responses, including induction of heat shock proteins (HSPs), can be therapeutic molecular targets for AF intervention [4].

Atrial remodelling

Electrical remodelling

In 1995, Wijffels et al. [5] demonstrated in a goat model that artificial maintenance of AF by implanted atrial pacemaker leads to marked shortening of the atrial effective refractory period (ERP), resulting in an increase in the rate, inducibility and stability of AF. This phenomenon was described as ‘AF begets AF’ [5]. The electrophysiological and molecular mechanisms underlying tachypacing-induced shortening of the atrial ERP have been explored [6–9]. In a report studying ionic currents in right atrial myocytes from dogs subjected to 1, 7, or 42 days of atrial pacing at 400/min, progressive increase in the duration of burst pacing-induced AF and reduced atrial ERP and adaptation to rate of the atrial ERP was associated with the progressive reduction in L-type Ca²⁺ channel current [7]. In the same canine model, atrial tachypacing caused time-dependent decreases in the ERP, conduction velocity, and wavelength which, along with increased regional heterogeneity, provided a substrate for AF [8]. The same research group [9] subsequently investigated intracellular Ca²⁺ transients and cell shortening in isolated, field-stimulated canine atrial myocytes, and demonstrated that tachypacing increases cellular Ca²⁺ loading, leading to post-tachypacing abnormalities in Ca²⁺ handling that produce contractile dysfunction. The reduction in the L-type Ca²⁺ channel caused by an increase in intracellular Ca²⁺ has been attributed to shortening of the atrial ERP during AF progression; a substrate for AF begets AF [10]. In the report investigating the relationship between gene expression of the L-type Ca²⁺ channel and the atrial ERP in patients with AF, the reduction in L-type Ca²⁺ channel gene expression correlated with shortening of atrial ERP [11]. Although several potassium channels and sodium channels may also be involved in tachypacing-induced atrial electrical remodelling [12], the main cause of atrial ERP shortening is thought to be due to reduction in L-type Ca²⁺ channel down-regulation (Figure 1) [13].

Figure 1. (A) A conceptual schema of the pathogenesis of atrial-tachycardia remodelling, leading to the development of atrial fibrillation (AF). As AF increases the heart rate by 10-fold, intracellular Ca²⁺ loading, a threat to cell viability occurs. In a short time (within minutes) the protective mechanism L-type Ca²⁺ channel current (I_Ca) is inactivated partly by intracellular Ca²⁺-dependent inactivation mechanism. Subsequently (hours to days), the expression of mRNA encoding L-type Ca²⁺ channel decreases followed by its protein expression reduction and I_Ca reduction, resulting in an abbreviation of both action potential duration (APD) and effective refractory period (ERP). Thus, the wavelength (WL), which is obtained by the product of conduction velocity and ERP, is reduced. (B) The number of waves that fit onto the atrial surface at any given time clearly depends on WL. (a) In a normal-size atrium with normal WL, only one wave can exist, thus sustained AF cannot be developed. (b) When WL is reduced by decreased ERP due to I_Ca down-regulation, as described in A, many waves can be accommodated in a same size of atrium, thus AF is likely to be sustained. (Figure adapted with permission [13]).

Structural remodelling

Atrial fibrosis is a common feature of AF [14]. Atrial tachypacing causes electrical remodelling characterized by atrial ERP shortening as described [5–11] but, in a canine AF model with rapid atrial pacing at 400 beats/min for 5 weeks, rather than shortening the atrial ERP, extensive interstitial fibrosis was found in the atrial free wall in association with gradual conduction prolongation in the atria [15]. The authors also found that long-term treatment with the angiotensin-II type-1 receptor (AT1R) blocker candesartan that started 1 week before pacing and continued for 5 weeks during tachypacing prevented atrial fibrosis. Goette et al. [16] demonstrated that activation of angiotensin-converting enzyme (ACE)-dependent extracellular signal-regulated kinases (Erk1/Erk2) is involved in the development of atrial fibrosis in AF patients. The authors also demonstrated that the atrial AT1R is down-regulated and AT2R is up-regulated in AF patients [17]. The structural remodelling characterized by atrial interstitial fibrosis caused by an activated renin–angiotensin system is therefore believed to participate in AF persistence. Treatment with the ACE inhibitor enalapril markedly reduces the risk of development of AF in patients with left ventricular dysfunction [18]. Treatment with the AT1R blockers candesartan and valsartan also reduced the incidence of AF in patients with heart failure [19],[20]. The effect of ACE inhibitors or AT1R blockers in terms of reduction in new-onset AF beyond the blood pressure-lowering effect in patients without structural heart disease is controversial [21].

Induction of HSPs on atrial remodelling

Four reports have described the effects of HSP induction in an experimental model of AF [22–25]. Brundel et al. [22] reported the effects of HSP up-regulation against tachypacing-induced myolysis. Using HL-1 atrial myocytes derived from mouse atria, they revealed that tachypacing (5 Hz)-induced myolysis, estimated by myosin disruption, was completely prevented by pre-treatment with hyperthermia or geranylgeranylacetone (GGA). Because these interventions induce many classes of HSP, the effects of transient transfection of human Hsp27 or human Hsp72 were evaluated. The authors found that Hsp27 but not Hsp72 transfection was sufficient for protection against tachypacing-induced myolysis. To test more directly the effects of HSP induction against atrial remodelling, Brundel et al. [23] subsequently demonstrated that tachypacing (3 Hz)-induced reduction in duration of L-type Ca²⁺ current and action potential was prevented in HL-1 cells exposed to GGA, and that protective effects require Hsp27 induction and phosphorylation. The authors also demonstrated in dogs in vivo that atrial tachypacing (400 beats/min for 7 days) reduced the atrial ERP and rate adaptation of ERP, which was attenuated by administration of GGA (120 mg/kg per day, p.o.). GGA also suppressed tachypacing-induced AF-promoting changes, including AF duration by burst pacing and AF vulnerability (percentage of atrial sites at which AF was induced by single premature stimuli). These observations suggest that HSP induction, particularly Hsp27, may be an effective approach to prevent progression of clinical AF.

Recent clinical evidence has suggested that clinical AF is frequently observed in patients with hypertension [1]. The stretch and/or activation of renin–angiotensin system causes atrial interstitial fibrosis, a substrate for AF progression [3]. Wakisaka et al. [24] tested the hypothesis that atrial fibrosis and AF evoked by angiotensin-II (AII) could be prevented by treatment with hyperthermia. In cultured rat left atrial fibroblasts, pre-treatment with hyperthermia (42°C for 30 min) induced Hsp72 expression and attenuated AII-induced Erk1/Erk2 phosphorylation, α-smooth muscle actin (α-SMA) expression, transforming growth factor-β1 (TGF-β1) secretion, collagen synthesis, and expression of collagen type-I and tissue inhibitor of metalloproteinases-1. Because a small interfering RNA targeting Hsp72 abolished these anti-fibrotic effects of hyperthermia, the authors concluded that heat-shock responses, particularly induction of Hsp72, have a dominant role in suppression of the AII-induced fibrotic signal. They also showed in rats in vivo that repeated hyperthermia (43°C, 20 min, once a week for 4 weeks) prevented the left atrial interstitial fibrosis induced by continuous infusion of AII, resulting in the reduced inducibility of AF by extrastimuli (Figure 2) [24]. When compared with the atrial tachypacing model, an approach using continuous infusion of AII without pacing may have some advantages. This is because continuous infusion of AII causes left atrial wall stretch (reflected as a slight increase in left ventricular end-diastolic pressure) without heart failure, and causes atrial interstitial fibrosis without affecting the atrial ERP. These processes are more likely to underlie the mechanisms for newly developed AF in patients with hypertension.

Figure 2. Histology of the left atrial (LA) free wall. Interstitial fibrosis was induced by continuous infusion of angiotensin II (AII: 400 ng/kg/min) by subcutaneously implanted osmotic mini-pump. Whole-body hyperthermia (HT: 43°C, 20 min) or normothermia (NT: 37°C, 20 min) was applied 24 h before and 7, 14, and 21 days after the start of the AII infusion. For HT treatment rats were anaesthetized with pentobarbital (20 mg/kg, IP) and placed, with their heads on a pillow to avoid aspiration of water, for 20 min in a bath in which the water temperature was maintained at 43°C. Rectal temperature was monitored throughout the thermal treatment experiments. Application of HT at 43°C elevated the rectal temperature to 41°C within 10 min and the temperature remained between 41° and 42°C during HT application. On day 28, LA was carefully removed, fixed, and stained with Masson trichrome. Blue staining represents interstitial fibrosis. To quantify interstitial fibrosis, the blue pixel content was measured relative to the total tissue area. A–C: Representative atrial tissue sections. Compared with the control group (A), blue-stained extensive and heterogeneous interstitial fibrosis was observed in the AII-NT group (B). The interstitial fibrosis was attenuated in AII-HT group (C). (Figure adapted with permission [24]).

Ischaemic heart disease is associated with an increased risk of AF; atrial ischaemia is a potential contributor to this association [1]. It has been shown in dogs that experimental atrial ischaemia produces conduction slowing in the ischaemic region, which stabilized the re-entrant circuit of AF [26]. Sakabe et al. [25] reported using the same atrial ischaemia dog model that Hsp72 induction by orally administered GGA-suppressed conduction heterogeneity and burst pacing-induced AF duration. In the ischaemic and non-ischaemic regions, Hsp72 was intensively induced by GGA, whereas Hsp27 induction was not significant. Studies from our research group demonstrated that Hsp72 induction attenuated ventricular ischaemia reperfusion injury [27–31]. We also reported that protective effects of Hsp72 induction against ventricular ischaemia reperfusion injury are observed even in diabetic rats [32],[33]. Hsp72 rather than Hsp27 therefore appears to have a more important role in the prevention of ischaemia-induced production of AF substrate.

Clinical evidence

AF after cardiac surgery is a frequent complication, and is associated with a longer stay in hospital and numerous postoperative complications [34],[35]. St Rammos et al. [36] investigated 101 patients who underwent open-heart surgery using cardiopulmonary bypass, and found that patients with low preoperative right atrial Hsp72 expression had a significantly greater incidence of postoperative AF. Mandal et al. [37] demonstrated, in 80 patients undergoing elective coronary artery bypass surgery, that preoperative Hsp72 content in right atrial tissue obtained at surgery was higher in patients who did not develop postoperative AF than in those who developed AF (Figure 3). These observations suggest that preoperative induction of Hsp72 as a measure of myocardial preconditioning may potentially reduce the incidence of postoperative AF. Regarding Hsp27, Brundel et al. [22] reported that expression of right and left atrial Hsp27 was significantly increased in patients with paroxysmal AF compared with those in non-AF control patients and patients with persistent AF. Because an inverse correlation was observed between duration of persistent AF and Hsp27 expression, the authors interpreted the observation that even though Hsp27 is essentially protective against AF, the persistence of AF exhausted the heat shock response, leading to a loss of its protective effects, thereby promoting progression to persistent AF. Taken together, it is suggested that the increased level of atrial expression of Hsp72 and Hsp27 may reflect their ability to prevent the progression of paroxysmal AF to persistent AF. It would be useful if the atrial tissue levels of HSP expression could be assessed by their serum concentration. Mandal et al. [37] observed a lack of correlation between atrial intracellular Hsp72 content and serum-soluble Hsp72 level, and that there was no association between serum-soluble Hsp72 and incidence of postoperative AF. The protective effects of Hsp72 may be exerted when it localizes intracellularly, and may be lost when it is released into blood. It is unclear whether serum HSP concentrations reflect intracellular levels, and further studies are needed to address these important issues.

Figure 3. Quantitative analysis of atrial Hsp72 content by Western blot analysis in 80 patients who underwent elective coronary artery bypass surgery. In each patient, a block of tissue was obtained from the right atrial appendage immediately after opening the pericardium. The Hsp72 content was compared between the patients who developed postoperative atrial fibrillation (AF) and those who did not develop postoperative AF. As shown, preoperative right atrial Hsp72 content was higher in patients who did not develop postoperative AF (AF (−)) than in those who developed AF (AF (+)). *p = 0.007). (Figure adapted with permission [37]).

Conclusion

Experimental studies have shown that atrial tachypacing, continuous exposure to AII (pressure overload) or acute atrial ischaemia causes electrical remodelling and structural remodelling, respectively. This is characterized by shortening of the atrial ERP and fibrosis-mediated conduction abnormality, leading to reduction in wavelength, which promotes AF likelihood (Figure 4). Induction of heat shock responses (including induction of Hsp72 and Hsp27 by hyperthermia and/or GGA) protects the heart against AF by preventing electrical remodelling and structural remodelling. The potent protective role of Hsp72 and Hsp27 against the progression of clinical AF is also evident. Taken together, interventions that induce heat shock responses may prevent newly developed AF and delay progression of paroxysmal AF to persistent AF.

Figure 4. Tachypacing, continuous infusion of angiotensin II (AII), and ischaemia induce electrical remodelling and structural remodelling in atria. This results in a reduction in wavelength, which is a substrate for the development and progression of atrial fibrillation (AF). Induction of HSPs can prevent atrial remodelling.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.