The heart is innervated by a sympathetic neuronal network that regulates its function, especially during conditions of increased physiological demand such as stress or exercise. Increased sympathetic activation is a hallmark of several cardiovascular diseases and arrhythmias. However, little is known about the physiological mechanisms underlying sympathetic control of the heart, hyperactivation or hyperinnervation.  Although significant progress has been made in understanding heart function, including the neurohormonal effects in the cardiomyocytes, understanding direct neuro-cardiac coupling mechanisms and the influence of neuronal density, contact and remodelling has been challenging. The aim of this study was to investigate neuro-cardiac interactions for the first time human primary cardiac neurons and myocytes .

Human atrial discarded tissue was obtained during cardiac surgery from 42 patients. Tissue samples were cleared of adipose tissue and only myocardium sections were used for cells isolation. Cells were then transduced with adenoviral vectors, encoding for the fluorescence resonance energy transfer (FRET) biosensor Epac1-camps, for real-time live-cell measurement of cytosolic cyclic adenosine monophosphate (cAMP)[1], and cultured during 48h as previously described[2]. Traditionally, neuronal somas are thought to be restricted to the central nervous system and the ganglia. However, we found a large population of functional resident neurons spread through the myocardium. We then co-cultured intracardiac neurons and myocytes and compared the FRET response to increasing concentrations of isoprenaline (1, 10 and 100 nM), a β-adrenergic agonist. Additionally, the cAMP-degrading capacity was assessed by measuring the response to 100 µM 3-isobutyl-1-methylxanthine (IBMX), a non-specific phosphodiesterase inhibitor, under isoprenaline stimulation and basal conditions. cAMP increase in response to 1 nM isoprenaline was not significantly different in isolated myocytes compared to those in contact with neurons (2.52 ± 0.63% vs. 6.40 ± 1.30%, p=0.126) but it was significantly greater at 10 nM (7.97 ± 1.63% vs. 14.36 ± 2.43%, p = 0.036) and 100 nM (14.08 ± 2.20% vs. 25.40 ± 4.20%, p=0.004). Importantly, cAMP increase to 1 nM isoprenaline in myocytes that were in contact to neurons was similar to the maximal response of 100 nM isoprenaline in myocytes alone. IBMX response was also stronger in myocytes in contact with neurons both in isoprenaline stimulation (21.58 ± 3.45% vs. 30.05 ± 5.09%, p=0.028) and basal conditions (15.39 ± 4.98% vs. 51.17 ± 5.97%, p<0.001), an effect that can only be explained by sympathetic neurotransmission. Normal distribution was confirmed by Shapiro-Wilk test. Data are expressed as mean ± standard error of the mean. Linear mixed models were used employing estimated marginal means as post-hoc test to obtain the p-values of pairwise comparisons.

These findings suggest that efficient sympathetic regulation of cardiac function occurs via direct coupling between myocytes and intracardiac neurons. The existence of this neuronal network and extensive cell-cell interactions could resolve the precise, specific and temporally tuning of heart function and provide the mechanistic framework for some cardiac arrhythmias.