Cardiac sympathetic innervation and arrhythmogenesis

The autonomic nervous system (ANS) plays an essential role in regulating many aspects of cardiac physiology including modulation of chronotropy, dromotropy, inotropy, and lusitropy (Gardner et al., 2016; Zaglia & Mongillo, 2017; Shivkumar, 2019). Dysregulation of the cardiac ANS, therefore, is involved in the pathogenesis of various cardiovascular diseases from hypertension, heart failure, myocardial infarction, to ventricular arrhythmia. Since sympathetic activation is believed to contribute to ventricular arrhythmogenesis, neuromodulation therapies for ventricular arrhythmia have targeted cardiac sympathetic innervation at different sites – β-blockers to inhibit β-adrenergic receptors, thoracic epidural anaesthesia to induce sympathetic block at the T1 to T4 levels, and stellate ganglion blockade with surgical or percutaneous approaches (Shivkumar, 2019). However, despite substantial progress in our understanding of the physiological function of the cardiac ANS, the dynamics and regulation of intercellular communication between cardiac sympathetic neurons (SN) and cardiomyocytes remain poorly understood.

Recent work from Drs. Zaglia and Mongillo’s groups at the University of Padova has shed novel insight into the molecular mechanisms by which cardiac SNs interact with and modulate ventricular cardiomyocytes (Zaglia & Mongillo, 2017). During acute distress, stimulation of cardiac SNs induces so called “fight-or-flight” response, which is associated with positive inotropic and chronotropic effects via activation of β-adrenergic receptors. Cardiac sympathetic innervation also provides basal and constitutive regulation of cardiomyocyte size and trophic signalling at resting conditions. Indeed, chemical sympathectomy experiments showed that denervation of cardiac SNs led to a significant reduction in the cardiac mass and cardiomyocyte size by affecting the β2-adrenergic receptor-mediated suppression of proteolysis/ubiquitin pathway (Zaglia et al., 2013). Direct intercellular communication in co-cultured SNs and cardiomyocytes has been investigated in vitro by studying live cell cAMP signalling in SN-sensitized cardiomyocytes in response to neuronal depolarization. The cAMP signalling elicited by β-adrenergic receptor activation is likely confined to a diffusion-limited domain at the intercellular contact site, reminiscent of the characteristics of the neuromuscular junction. Furthermore, spatially selective activation of cardiac SNs in vivo with optogenetics to modulate sinoatrial node function suggests direct coupling of SNs and effector cells, supporting the hypothesis of locally released noradrenaline at a junctional intercellular domain (Prando et al., 2018).

To further delineate the physiological coupling between cardiac SNs and cardiomyocytes, Pianca et al recently performed well-designed experiments to map the topology of cardiac SN distribution and examine the relationship between SN innervation and transmural heterogeneity in cardiomyocyte size with correlation to specific intercellular neuro-cardiac junctions (Pianca et al., 2019). Firstly, they showed that in the in vitro co-culture system that was maintained for 15 days, cardiomyocytes innervated by SNs displayed a significant 1.4-fold increase in sizes compared with the non-innervated counterparts (cardiomyocyte area, innervated: 1197.03 ± 84.46 vs. non-innervated: 844.47 ± 59.38 μm², respectively), consistent with prior observation that cardiac SNs elicit trophic inputs on innervated cardiomyocytes (Zaglia et al., 2013). In the transmural (short-axis) analysis of adult rodent hearts quantified by the number of tyrosine hydroxylase-positive fibres per cardiomyocyte, the sub-epicardium (EPI) layer contained three times denser SN innervation than the sub-endocardium (ENDO) layer, which correlated with larger cardiomyocyte cross-sectional areas in the EPI cardiomyocytes compared with the ENDO cardiomyocytes. In contrast, there was no significant difference in the longitudinal axis in terms of cardiomyocyte length.

Secondly, to overcome the limitation of thin-sliced tissue samples in a two-dimensional structure, the three-dimensional SN network within the intact anatomical context was stained by an anti-tyrosine hydroxylase antibody and was visualized by using the whole-mount two-photon immunofluorescence characterization of tissue-clarified murine myocardial blocks. Nearly all cardiomyocytes across the left ventricular wall were innervated and more SN processes per cardiomyocyte were detected in the EPI region compared with the ENDO region (1.69 ± 0.10 vs. 1.24 ± 0.08, respectively). Similar complex myocardial innervation network with highly arborized, regularly distributed varicosities was also found in a post-mortem human heart.

Next, to decipher the direct in vivo effects of cardiac SNs on cardiomyocyte size, Pianca et al showed that when the cardiac SN innervation gradually reached completion around the 3rd postnatal week at P21, the ratio between the cross-section areas of EP and ENDO cardiomyocytes was also progressively increased through the postnatal period from the baseline ratio ≅1 at P1 when SN innervation was absent (Pianca et al., 2019). This SN innervation effect on cardiomyocyte size was abolished by pharmacological sympathectomy to disrupt early cardiac SN innervation, causing cardiac atrophy. Similarly, pharmacological neuronal ablation in adult mice, not caloric restriction, resulted in diminished transmural heterogeneity in cardiomyocyte size. In addition, since cardiac SN-induced trophic input relies on the β2-adrenergic receptor/ubiquitin ligase MuRF-1 signalling axis , modulation of the signalling axis with either pharmacological inhibition (atrophic effect) or overstimulation (hypertrophic effect) of β2-adrenergic receptors, or gene knockout of MuRF1, resulted in cardiac remodelling and abolished transmural differences without cardiac SN innervation disruption.

Finally, because SN denervation causes a more dramatic change in the EPI myocardium to diminish the transmural differences, MuRF1 expression level and activity were investigated after chemical denervation. Indeed, MuRIF1 expression level was significantly higher in the EPI region than in the ENDO region after denervation. MuRIF1-directly ubiquitination of cardiac troponin I was also increased in the EPI myocardium. These findings indicate that transmural heterogeneity of cardiomyocyte size mediated by the distinct cardiac SN distribution across the ventricular wall is associated with local regulation of the cardiomyocyte proteolytic machinery.

Taken together, Pianca et al has provided novel physiological aspects of differential cardiac SN innervation on the ventricular structural remodelling at the molecular, cellular and organ levels (Pianca et al., 2019). How can these novel findings be translated into our understanding of pathophysiological mechanisms of ventricular arrhythmia? In heart transplant patients, SN denervation occurs immediately following surgical interruption of the SN network. Sympathetic reinnervation, if occurring later on, appears to be incomplete and displays heterogenous patterns on the ventricular wall. Theoretically, interruption of SN innervation would lead to impaired transmural heterogeneity of cardiomyocyte size because of loss of trophic signalling through cardiac SNs. Partial and heterogenous restoration of sympathetic innervation might cause variations across the ventricular surface area and transmural regions in the ventricle, which, in return, could exaggerate arrhythmogenic response to stimuli such as catecholamines.

As shown in Fig. 2A (Pianca et al., 2019), Kv4.2 is differentially expressed between EPI and ENDO myocardium. Could SN denervation and loss of trophic signalling also affect ion channel expression besides reducing cardiomyocyte size? The answer is probably yes as demonstrated in prior cardiac K+ channel studies. Regional variations in sympathetic dysinnervation through either hyperinnervation or denervation likely contribute to serious ventricular arrhythmia and sudden cardiac death in myocardial infarction and heart failure (Gardner et al., 2016; Shivkumar, 2019). β-blockers remain the first-line anti-arrhythmic therapy for ventricular arrhythmia, especially in patients with heart failure with reduced ejection fraction and polymorphic ventricular tachycardia after myocardial infarction (Gardner et al., 2016; Al-Khatib et al., 2018).

In ventricular tachycardia/fibrillation storm refractory to anti-arrhythmic medications and catheter ablation, cardiac SN denervation by transient chemical approach (thoracic epidural anaesthesia) or permanent surgical approach (stellate ganglionectomy) has been used to manage these life-threatening conditions (Shivkumar, 2019). Due to technical difficulties and limited data available from clinical studies (Al-Khatib et al., 2018), cardiac SN denervation has not been widely applied in clinical settings. The basic science behind SN denervation is also not comprehensive given little is known about how the ANS is remodelled during the initiation and progression of ventricular arrhythmia. The elegant studies undertaken by Pianca et al (Pianca et al., 2019), together with their prior reports on dynamic intercellular communication between cardiac SN innervation and cardiomyocytes (Zaglia et al., 2013; Zaglia & Mongillo, 2017; Prando et al., 2018), will no doubt help further advance our understanding and ability to determine the molecular mechanisms of sympathetic remodelling-induced ventricular arrhythmia and identify potential new therapeutic targets.