Dysregulation of Neuronal Ca2+ Channel Linked to Heightened Sympathetic Phenotype in Prohypertensive States

Hypertension is associated with impaired nitric oxide (NO)–cyclic nucleotide (CN)-coupled intracellular calcium (Ca2+) homeostasis that enhances cardiac sympathetic neurotransmission. Because neuronal membrane Ca2+ currents are reduced by NO-activated S-nitrosylation, we tested whether CNs affect membrane channel conductance directly in neurons isolated from the stellate ganglia of spontaneously hypertensive rats (SHRs) and their normotensive controls. Using voltage-clamp and cAMP–protein kinase A (PKA) FRET sensors, we hypothesized that impaired CN regulation provides a direct link to abnormal signaling of neuronal calcium channels in the SHR and that targeting cGMP can restore the channel phenotype. We found significantly larger whole-cell Ca2+ currents from diseased neurons that were largely mediated by the N-type Ca2+ channel (Cav2.2). Elevating cGMP restored the SHR Ca2+ current to levels seen in normal neurons that were not affected by cGMP. cGMP also decreased cAMP levels and PKA activity in diseased neurons. In contrast, cAMP–PKA activity was increased in normal neurons, suggesting differential switching in phosphodiesterase (PDE) activity. PDE2A inhibition enhanced the Ca2+ current in normal neurons to a conductance similar to that seen in SHR neurons, whereas the inhibitor slightly decreased the current in diseased neurons. Pharmacological evidence supported a switching from cGMP acting via PDE3 in control neurons to PDE2A in SHR neurons in the modulation of the Ca2+ current. Our data suggest that a disturbance in the regulation of PDE-coupled CNs linked to N-type Ca2+ channels is an early hallmark of the prohypertensive phenotype associated with intracellular Ca2+ impairment underpinning sympathetic dysautonomia. SIGNIFICANCE STATEMENT Here, we identify dysregulation of cyclic-nucleotide (CN)-linked neuronal Ca2+ channel activity that could provide the trigger for the enhanced sympathetic neurotransmission observed in the prohypertensive state. Furthermore, we provide evidence that increasing cGMP rescues the channel phenotype and restores ion channel activity to levels seen in normal neurons. We also observed CN cross-talk in sympathetic neurons that may be related to a differential switching in phosphodiesterase activity. The presence of these early molecular changes in asymptomatic, prohypertensive animals could facilitate the identification of novel therapeutic targets with which to modulate intracellular Ca2+. Turning down the gain of sympathetic hyperresponsiveness in cardiovascular disease associated with sympathetic dysautonomia would have significant therapeutic utility.


Introduction
Increased activity of the sympathetic nervous system is a well established contributing factor to the pathophysiology of human hypertension (Esler, 2010; and to that observed in the spontaneously hypertensive rat (SHR) (Heaton et al., 2006;Shanks et al., 2013a;Mancia and Grassi, 2014). In par-ticular, the renal nerves (Lundin et al., 1984) and the postganglionic nerves innervating the heart are sympathetic beds associated with heightened activity. Interestingly, this sympathetic impairment occurs in young, prohypertensive animals Shanks et al., 2013c), suggesting that it is an early hallmark of the disease itself, although the exact mechanisms that underpin the phenotype remain poorly understood. When postganglionic sympathetic stellate neurons (PGSNs) that predominantly innervate the heart from age-matched prohypertensive SHRs are compared with PGSNs from normotensive rats (WKY), the SHR neurons have increased intracellular calcium transients ([Ca 2ϩ ] i )  and enhanced neurotransmission (Zugck et al., 2003;Li et al., 2007;Shanks and Herring, 2013;Shanks et al., 2013c). This translates into increased postsynaptic excitability of the sinoatrial node due to a greater conductance of ICa L 2ϩ and an increase in heart rate (Heaton et al., 2006). The observed differences in [Ca 2ϩ ] i may be partially explained by altered mitochondrial Ca 2ϩ handling , but it is unclear whether there are additional changes at the membrane level that contribute to the greater calcium transients.
Recently, Tu et al. (2014) demonstrated enhanced activity of the N-type voltage-gated Ca 2ϩ (Cav 2.2 ) channel in the PGSNs of a rat model of heart failure. Another study found that inhibition of this channel restores cardiac autonomic imbalance and prevents lethal arrhythmias in a mouse model of heart failure (Yamada et al., 2014). Similarly, we also observed increased whole-cell neuronal Ca 2ϩ currents in the prohypertensive SHR , suggesting that channel dysregulation is an early molecular event that may be highly conserved over several developmental stages from the prediseased phenotype to advanced heart failure (Li and Paterson, 2016). What links ion channel dysregulation to sympathetic hyperresponsiveness? One suggestion is that early metabolic-oxidative stress could be coupled to abnormal cyclic nucleotide (CN) regulation (Krukoff, 1998;Danson and Paterson, 2005;Danson et al., 2005). Voltagegated Ca 2ϩ channels are strongly regulated by CN-coupled protein kinases (Catterall, 2000;Simms and Zamponi, 2014) that are dysfunctional in hypertension Heaton et al., 2007;Li et al., 2013). Specifically, PGSNs have reduced neuronal nitric oxide synthase (nNOS) activity, low level expression of soluble guanylyl cyclase (sGC)  and high phosphodiesterase 2A (PDE2A) activity , leading to reduced levels of cGMP . Together, this cascade of events could provide a link to ion channel dysregulation. However, emerging evidence suggests that the inhibitory action of NO donors on Cav 2.2 in HEK 293 cells is via cGMP-independent pathways coupled to S-nitrosylation of the N-type channel (Jin et al., 2011;Zhou et al., 2015).
Here, we tested the hypothesis that impaired CN regulation provides a direct link to abnormal signaling of neuronal calcium channels in prohypertensive rats and that targeting cGMP can restore the normal channel phenotype in the diseased PGSN.

Materials and Methods
Animals. Four-week-old male prohypertensive SHRs and normotensive WKY and Wistar (hereafter referred to as control) rats were obtained from Harlan Laboratories and Charles River Laboratories. The SHRs begin to show development of clinical symptoms of hypertension at 6 weeks of age (Dickhout and Lee, 1998). The rats were housed in standard plastic cages and artificial lighting was fixed to a natural 12 h light/dark cycle. Food and water were available ad libitum. All experiments were performed in accordance with the UK Home Office Animals Scientific Procedures Act of 1986 (PPL 30/2360 and 30/3131, David J. Paterson).
Reagents. All drugs and reagents were sourced from Sigma-Aldrich unless stated otherwise.
Adenovirus generation. The complete open reading frame of the cAMP-sensitive FRET-based sensor H187 was obtained from pcDNA3 using double digestion with the restriction enzymes EcoRI and HindIII and introduced in the pDual vector. Generation and amplification of the viral particles were outsourced to Vector Biolabs.
Cell isolation. The sympathetic stellate neurons that predominately innervate the heart were isolated as described previously (Shanks et al., 2013b). Briefly, the stellate neurons from 4-week-old prohypertensive SHR and normotensive control rats were dissected out and digested with collagenase (1 mg/ml) and trypsin (2 mg/ml) to create a single-cell suspension. The PGSNs were then grown on 6 mm poly-D-lysine/laminincoated coverslips (37°C 5% CO 2 ) for 3 h before electrophysiological recordings were made. For measurements of intracellular cAMP and protein kinase A (PKA), PGSNs were transfected with relevant adenoviral particle-forming units (H187:1.4 ϫ 10 7 , AKAR4: 4 ϫ 10 9 ) for 24 h. The cells were incubated for 24 -48 h before FRET imaging.
Whole-cell voltage clamp of acutely isolated neurons. Voltage-clamp experiments were conducted using the Axopatch 200B amplifier and Clampex software. The neurons were held at Ϫ90 mV before a set of 50 ms voltage steps from Ϫ50 to ϩ50 were applied with 10 mV, 5 s intervals. The bath was grounded using an Ag/AgCl silver pellet wire connected to the bath via a 3 M agar salt bridge. Cells were perfused with tyrode or Ca 2ϩ -isolating solution using a custom-built perfusion system into a heated microscope stage chamber maintained at 37°C. Access to the cell was obtained in normal tyrode solution containing the following (in mM): 135 NaCl, 4.5 KCl, 11 glucose, 20 HEPES, 1MgCl 2 , 2 CaCl 2, pH 7.4. Then, the solution was changed to the Ca 2ϩ -isolating solution containing no Na ϩ and K ϩ channel blockers (TEA and Cs ϩ ) with Ba 2ϩ as the charge carrier to isolate the Ca 2ϩ currents containing the following (in mM): 135 TEACl, 10 HEPES, 4.5 KCl, 1 MgCl 2 , 4 glucose, 1 NaHCO 3 , 2 Ba 2ϩ , pH 7.40, either in the presence or absence of drugs (1 M -conotoxin GVIA or 100 M 8 bromoguanosine-cGMP, 8b-cGMP). Ba 2ϩ was used as the charge carrier to avoid Ca 2ϩ -dependent current inactivation (Imredy and Yue, 1994). The internal solution contained the following (in mM): 140 CsCl, 10 HEPES, 0.1 CaCl 2 , 1 MgCl 2 , 4 MgATP, 1 EGTA, pH 7.30. All solutions had osmolarities of 300 mOsm/L. A 75-90% series resistance compensation was applied to all cells. Peripheral neurons portray significant axonal branching, making an effective voltage clamp difficult to achieve (Bar-Yehuda and Korngreen, 2008). To minimize these space-clamp errors, the cells were held at Ϫ90 mV, low concentrations of the charge carrier was used, cells were patched acutely, and large bore, blunt electrodes (1.5-2 M⍀) were used in addition to strict exclusion criteria. Only cells that had a series resistance of Ͻ10 M⍀ throughout the recording, a holding potential of ϾϪ300 pA, and cells in which no visible contusion of the currents were visible were included in this study. Due to significant run-down of the currents in both animal models, which was found to be neither time nor activity dependent, unpaired experiments were performed to unmask differences between cells.
FRET measurements of cytosolic cAMP and PKA activity. For intracellular measurements of cAMP and PKA activity, cells were infected 0 -24 h after culture with the FRET-based cAMP sensors Epac-H30 (Ponsioen et al., 2004), Epac-H187 (Klarenbeek et al., 2015) or the PKA sensor AKAR4 (Depry et al., 2011; Lefkimmiatis et al., 2013) for 24 h. FRET sensor-expressing neurons were imaged 3-4 d after isolation on an inverted Nikon microscope. This was connected to an OptoLED fluores-cence imaging system (Cairn Research ) equipped with a 40ϫ oilimmersion objective, a CoolSnap HQ2 digital CCD camera (Photometics), and a beam-splitter (DV2; Photometrics), which included the emission filters for CFP and YFP acquisition (dichroic mirror 505DCXR). The cells were excited at 430 nm for 100 ms every 15 s. CFP and YFP emissions were measured and the ratio between 480/535 nm fluorescent emission intensities were automatically calculated using Optofluor software. Background fluorescence was subtracted from emission intensities and intensity ratios were plotted against time. Mean FRET responses were expressed as the percentage change from baseline (⌬R/R 0 where ⌬ R ϭ R Ϫ R 0 . R 0 is the ratio of intensity over a 30 s period in the absence of drugs and R is the mean ratio over 30 s in the presence of the respective drug treatment).
During FRET experiments, cells were perfused continuously with Tyrode's solution and flow rate was controlled at 2-3 ml/min. Phar-macological agents were diluted in Tyrode's solution and perfused at the following concentrations: forskolin, 0 -25 M; 3-isobutyl-1methylxanthine (IBMX), 1-100 M; the cGMP analog 8b-cGMP, 100 M; the PDE3 inhibitors cilostamide, 10 M, or milrinone, 10 M; and the PDE2 inhibitor BAY-60-7550, 1 M (Cayman Chemicals). For comparisons between cells, the average percentage FRET change over a 30 s period was calculated once equilibrium was reached. In all experiments, the maximal FRET change of each cell was recorded by exposing the cells to saturating concentrations of an adenylyl cyclase (AC) activator and a PDE inhibitor (25 M forskolin and 100 M IBMX, respectively) to ensure that the cells responded similarly to the sensor. The H30 cAMP sensor responded differently in the SHR and control cells, so these data were normalized to the IBMX/forskolin maximum FRET response to allow for comparisons between the control and SHR neurons. Figure 1. The whole-cell Ca 2ϩ current is larger in the prohypertensive SHR. Whole-cell voltage clamp was performed on the cardiac sympathetic stellate ganglion innervating the heart to investigate the whole-cell Ca 2ϩ properties of 4-week-old prohypertensive SHR and normotensive control rats; 50 ms, 10 mV voltage steps from Ϫ50 to ϩ50 were applied to the cell before the resulting current was measured. Immunofluoresence showed TH positivity, confirming sympathetic phenotype of the neurons (A). B, Representative voltage-clamp traces of control and SHR neurons showing significantly larger currents in the SHR neurons. C, Peak current (Ϫ10 mV) was Ϫ127.5 Ϯ 5.94 pA/pF (n ϭ 10) in the SHR and Ϫ108.0 Ϯ 6.80 pA/pF (n ϭ 10, p Ͻ 0.045) in the control. D, Current-voltage relationship showing significantly larger currents in the SHR compared with the control at multiple voltages. E, There were no changes in the voltage dependency of the channel, as shown by the activation curve, suggesting that the biophysical properties of the channels are unaffected. Data are represented as the mean Ϯ SEM.
Protocols. Specifically, we looked at the cells' ability to generate cAMP and resulting PKA activity by administering the AC activator forskolin. Further, we assessed the cells ability to hydrolyze cAMP by pharmacologically inhibiting the predominant PDE subtypes (PDEs 1-7, 10 -11) with the nonspecific PDE inhibitor IBMX. To test the involvement of the CNs in the regulation of the N-type channel, 8b-cGMP was perfused before the voltage-clamp experiment was repeated. In addition, to investigate whether the cGMP dysfunction seen in the prohypertensive state interferes with the cAMP pathway (through cross-talk via PDEs), we administered 8b-cGMP to the control and SHR neurons and monitored the resulting cAMP levels and PKA activity. Finally, to test the relative involvement of PDE2A and PDE3 in the cells, the cells were incubated with selective PDE2 (1 60 -7550) and PDE3 inhibitors (cilostamide or milrione) while the cAMP levels were monitored.
Statistical analysis. All data were analyzed using GraphPad Prism software. When the data passed normality tests, unpaired two-tailed t tests were used; when they did not, nonparametric tests were used with the specific test reported in the figure legend. All data are expressed as the mean Ϯ SEM. Statistical significance was accepted at p Ͻ 0.05.

Results
Neuronal Ca 2؉ currents of the prohypertensive SHR are larger than that of the normotensive control Immunofluorescence analysis of the cardiac stellate neurons confirmed their sympathetic phenotype by their TH positivity (Fig.  1A). The whole-cell voltage-gated Ca 2ϩ currents of the prohypertensive SHR (Ϫ127.5 Ϯ 5.94 pA/pF, n ϭ 10) were significantly larger than that of the normotensive control animals (Ϫ108.0 Ϯ Figure 2. 1 M conotoxin GVIA blocks 75% and 83% of the current, suggesting that the N-type Ca 2ϩ channel is the major contributor to ionic conductance. A, Immunofluorescence showing TH positivity and cell surface expression of the N-type Ca 2ϩ channel. The nuclear stain DAPI is included on the overlay image. Representative peak (Ϫ10 mV) voltage-clamp traces of control (B) and SHR (C) in normal conditions (solid lines) and after treatment with conotoxin GVIA (dashed lines). D, Current-voltage relationship showing uniform reduction in both the SHR and control neurons at multiple voltages. The peak current remains at Ϫ10 mV and is not significantly different between the control and SHR. E, Bar chart of the peak currents (Ϫ10 mV) showing a 75% (control) and 83% (SHR) reduction of the N-type Ca 2ϩ current down to levels that were not significantly different between the strains (Ϫ26.88 Ϯ 1.7 pA/pF, n ϭ 6 and Ϫ22.04 Ϯ 1.60 pA/pF, p ϭ 0.07 n ϭ 5). Dashed lines represent the mean of the control (black) and SHR (red) control data. Data are represented as the mean Ϯ SEM.
6.80 pA/pF, n ϭ 10, p Ͻ 0.045, unpaired t test) at multiple voltages ( Fig. 1B-D). No significant change in the activation curve was observed (Fig. 1E), suggesting that the biophysical properties of the channels were unchanged. No significant difference was observed in cell size (as measured by the cell capacitance in picofarads) or resting membrane potential between the control or SHR neurons (Ϫ46.75 Ϯ 0.95 vs Ϫ45.40 Ϯ 0.87 mV and 21.36 Ϯ 1.03 vs 21.32 Ϯ 1.17 pF, n ϭ 32 and 30, unpaired t test).
Increased whole-cell Ca 2؉ currents are attributed to changes in Cav 2.2 Immunofluorescence analysis confirmed the expression of the N-type Ca 2ϩ channels on the membrane of the TH-positive cells ( Fig. 2A). We tested their functional relevance by perfusing the cells with 1 M -conotoxin GVIA and observed a reduction in the whole-cell currents by 75% and 83%, respectively, down to similar levels in both normotensive and pro-hypertensive phenotypes (control Ϫ26.88 Ϯ 1.70, n ϭ 6; SHR Ϫ22.04 Ϯ 1.60 pA/pF, n ϭ 5, p ϭ 0.072), suggesting that Ca v2.2 is the Ca 2ϩ channel predominantly carrying the Ca 2ϩ current in PGSNs (Fig. 2B-E).

Increasing the intracellular cGMP concentrations significantly lowers Ca 2؉ currents and reverses the channel phenotype
To test the involvement of the CNs in the regulation of the N-type channel, the cell-permeant 8b-cGMP was perfused before the voltage-clamp experiment was repeated. Elevating intracellular cGMP did not alter the Ca 2ϩ currents in the normotensive control animals, but reduced the SHR currents (Ϫ127.5 Ϯ 5.94 pA/ pF, n ϭ 10 to Ϫ105.2 Ϯ 7.79 pA/pF, n ϭ 7, p ϭ 0.035) down to levels seen in the control animals (Ϫ108.0 Ϯ 6.80 pA/pF, n ϭ 10, p ϭ 0.79; Fig. 3A-E). No changes were observed in the activation curve (Fig. 3F ). Restoring the cGMP levels rescues the channel phenotype seen in the SHR. Representative peak voltage-clamp traces at Ϫ10 mV of control (A) and SHR (B) neurons showing a decrease in the SHR current after application of 100 M of a membrane-permeable cGMP analog 8b-cGMP. Current-voltage relationship of control (C) and SHR (D) neurons shows a normalization of the SHR current down to control levels at all voltages. The peak current remains at Ϫ10 mV and is not significantly different from the controls. The drug does not significantly alter the control currents. E, Bar chart of the peak currents (Ϫ10 mV) showing a significant reduction and normalization of the SHR current (Ϫ127.5 Ϯ 5.94 pA/pF, n ϭ 10 to Ϫ105.2 Ϯ 7.79 pA/pF, n ϭ 7, p ϭ 0.035) down to control levels (Ϫ108.0 Ϯ 6.80 pA/pF, n ϭ 10, p ϭ 0.79). Dashed lines represent the mean of the control (black) and SHR (red) control data. F, Activation plot showing no significant difference between the SHR and control or the drug groups, suggesting that the biophysical properties of the channels are unchanged. Data are represented as the mean Ϯ SEM.

Differential cGMP-cAMP cross-talk in the SHR
To investigate whether the cGMP dysfunction seen in the prohypertensive state was interfering with the cAMP pathway (through cross-talk via PDEs), we administered 8b-cGMP to the neurons and monitored cAMP levels. We then established the functional relevance of this by assessing PKA activity. Application of 8b-cGMP led to a decrease in cAMP levels (Ϫ4.82 Ϯ 2.02%, n ϭ 14), with no change in PKA activity (1.09 Ϯ 0.57%, n ϭ 8) in the SHR neurons (Fig. 4 A, B). In contrast, when normotensive neurons were treated with 8b-cGMP, we observed a significant increase in both cAMP levels (11.68 Ϯ 1.54% n ϭ 16) and PKA activity (19.15 Ϯ 3.51%, n ϭ 6; Fig. 4C,D). This finding suggests a differential involvement of PDE isoforms in response to cGMP signaling in the normotensive versus prohypertensive PGSNs.
Differential response of the SHR and control neurons to PDE2A blockade PDE2A has been shown recently to play a major role in the control of neurotransmission in sympathetic neurons. To assess whether PDE2A plays a role in regulating Ca 2ϩ channels, we bath applied the specific PDE2A inhibitor BAY 60 -7550 and repeated the voltage-clamp experiments. The application of 1 M BAY 60 -7550 significantly increased the channel currents (Ϫ108.0 Ϯ 6.803 pA/pF, n ϭ 9 to Ϫ138.7 Ϯ 9.610 pA/pF, n ϭ 10, p ϭ 0.0169) in the normotensive neurons. Interestingly, the SHR neurons responded to the same treatment with a slight, nonsignificant decrease of currents (Ϫ127.5 Ϯ 5.937 pA/pF, n ϭ 10 to Ϫ118.0 Ϯ 6.673 pA/pF, n ϭ 9). After PDE2A inhibition, the control currents were trending toward being larger than the SHR, but this was not quite significant (138.7 Ϯ 9.610 pA/pF to Ϫ118.0 Ϯ 6.673 pA/pF, p ϭ 0.052; Fig. 5A-C).

cAMP-PKA signaling pathway remains unchanged in prohypertensive animals
To determine the cells' ability to generate cAMP and activate PKA, we administered the AC activator forskolin and measured the FRET responses of the two sensors. FRET experiments in which the PGSNs were challenged with increasing doses of forskolin (0.1-25 M) revealed no significant difference between the generation of cytosolic cAMP (n ϭ 15-18/21-28), and PKA activity (n ϭ 20 -23/16 -17) between 4-week normotensive control and pro-SHR, respectively (Fig. 7 A, B). This indicates that the cAMP-PKA signaling pathway remains intact in the prohypertensive animals.

Discussion
The novel findings in this study are as follows. The enhanced whole-cell Ca 2ϩ current from the prohypertensive SHR PGSNs are predominately mediated by activation of the N-type Ca 2ϩ channel. Elevating cGMP restored the SHR Ca 2ϩ current to levels seen in normal neurons. Control neurons were unaffected by raised cGMP. Elevated cGMP also decreased cAMP levels and PKA activity in diseased neurons. In contrast, cAMP-PKA activity was increased in normal neurons, suggesting differential switching in PDE activity. PDE2A inhibition enhanced the Ca 2ϩ current in normal neurons to a similar conductance to that seen in SHR neurons, whereas the inhibitor slightly decreased the current in diseased neurons. Global cAMP-PKA levels were not different between neurons when activated by an adenylyl cyclase activator. However, pharmacological evidence suggests a switch-ing from cGMP acting via PDE3 in the control neurons to PDE2A in the SHR neurons in the modulation of the N-type Ca 2ϩ current.

Channel dysfunction in cardiovascular diseases
The importance of the N-type Ca 2ϩ channel function in relation to the release of noradrenaline (NA) in the sympathetic nervous system has long been known (Molderings et al., 2000;Uhrenholt and Nedergaard, 2005), but its relevance to hypertension came from the observation that N-type channel blockade causes hypotension (Pruneau and Bélichard, 1992). Moreover, the Cav 2.2 mutant mouse has significant problems with blood pressure control that is linked to reduced baroreflex control of heart rate Mori et al., 2002). Recently, it was shown that stellate neurons from rats with from heart failure have markedly increased N-type channel currents (Tu et al., 2014). Interestingly, the increased current was not correlated with channel expression, suggesting a role for posttranslational regulation of channel activity. Although the N-type channel is the major contributor of Ca 2ϩ entry coupled to sympathetic transmission (Uhrenholt and Nedergaard, 2003), it is possible that other Ca 2ϩ channels are also involved and could account for the residual response that we observed after -conotoxin GVIA treatment. Although -conotoxin GVIA has high affinity for the N-type channel (Olivera et al., 1987), in rat sympathetic neurons, this drug has also been shown to block L-type currents (McCleskey et al., 1986). It is therefore possible that the effect that we observed could be attributed partially to overactivity of this channel in the SHR (Heaton Figure 7. cAMP-PKA signaling pathway is functional in SHR neurons. To investigate the ability of the cells to generate cAMP and PKA, activity was compared between the SHRs and controls using FRET imaging. Cells were infected with the H30 Epac and AKAR sensors and the changes in fluorescence emission intensity of CFP and YFP was monitored (A). No significant differences were found in forskolin-induced cAMP generation (B) or PKA activity (C) in the sympathetic neurons from the prohypertensive SHR or controls. Comparisons were made using the Mann-Whitney test. et al., 2006) because its inhibition on the mesenteric artery significantly reduces NA release in response to neuronal stimulation (Tsuda et al., 2001). Another possibility is that alterations in cell excitability could be linked to N-type Ca 2ϩ channel activity. Therefore, we cannot rule out that such changes in threshold or action potential frequency might also be linked with the N-type Ca 2ϩ channel phenotype. This could conceivably contribute to the overall modulation of channel excitability and enhanced neurotransmission.

CN regulation of Ca 2؉ channels
Ca 2ϩ channels are regulated by CNs via PKA-and protein kinase G (PKG)-dependent phosphorylation. cAMP and cGMP act via PKA and PKG to cause activation and inhibition of the channel current, respectively (Bannister et al., 2005;Catterall, 2011;Zamponi et al., 2015). This occurs either via direct phosphorylation of the channels (Tohse and Sperelakis, 1991;Jiang et al., 2000) or through cross-play between the two CN pathways via PDEs (Conti and Beavo, 2007;Zaccolo and Movsesian, 2007;Stangherlin et al., 2011;Stangherlin and Zaccolo, 2012a). CNs also act to terminate their own signals by activating protein phosphatases that dephosphorylate downstream targets and PDEs that break down cAMP and cGMP (Francis et al., 2010, 2011). The larger Ca 2ϩ currents observed in the SHR neurons could be due to reduced PKG-mediated inhibition and/or increased PKA-mediated activation of the current. Because -conotoxin GVIA reduced the currents of both models down to similar levels, this suggests that the differences seen are due to modulation of the N-type Ca 2ϩ channel. Moreover, we observed that currents were normalized after a 4-minute perfusion of cGMP, suggesting that changes in channel regulation rather than expression are likely to account for the differences reported here.

NO-cGMP pathway in sympathetic responsiveness
SHR sympathetic neurons have lower bioavailability of NO and reduced expression of the ␤ 1 subunit of sGC , resulting in lower levels of cGMP . This is associated with increased Ca 2ϩ transients  and enhanced neurotransmission Li et al., 2012;Shanks et al., 2013c) compared with control neurons. Gene transfer of nNOS  or its adaptor protein, CAPON , rescues this sympathetic phenotype. However, we cannot rule out that a part of the inhibitory effect of NO modulation on Cav 2.2 is via non-GMP-mediated S-nitroyslation of the channel protein itself. Zhou et al. (2015) reported in HEK 293 cells that the NO donor SNAP can generate S-nitrosothiols that block the N type conductance, an effect reversed by methanethiosulfonate ethylammonium. Moreover, they identified the consensus motifs for channel inhibition that involved several intracellular cysteine residues. In particular, site-directed mutagenesis Figure 8. Sympathetic neurotransmission. The arrival of an action potential in the synaptic nerve terminal leads to Ca 2ϩ entry via voltage-gated Ca 2ϩ channels. This triggers exocytosis of noradrenaline into the synaptic clefts, which acts on the ␤ 1 adrenoreceptor, altering myocyte behavior. In hypertension, nNOS expression and the ␤ 1 subunit of sGC expression is down, resulting in lower levels of cGMP, which, coupled with enhanced PDE2A activity, results in increased neurotransmission. Pharmacological elevation of cGMP reverses the channel phenotype, but also leads to changes in cAMP levels and PKA activity as a result of interactions with PDEs. The differential responses of the control and SHR neurons to cGMP elevation suggests that cAMP levels are controlled by both PDE2A and PDE3 subtypes in the control neurons, but that the prohypertensive phenotype is associated with PDE2A dominance. of cysteine 1045 in the II-III loop significantly reduced NO sensitivity of Cav 2.2 . Whether this pathway is functionally activated within the range of physiological concentrations of NO remains to be established.
Nevertheless, our results provide direct evidence in sympathetic neurons that cGMP causes attenuation of the Ca 2ϩ channel current. Indeed, previous studies have demonstrated that the NO-cGMP pathway can regulate Ca 2ϩ currents via PKG activation, but also through cross-talk with PDEs (Méry et al., 1991;Fischmeister and Méry, 1996). Studies have reported that application of 8b-cGMP leads to a direct phosphorylation of Ca 2ϩ channels via PKG (Jiang et al., 2000). This resulted in smaller currents, but no change in the voltage dependency (Tohse and Sperelakis, 1991), of the channel, which is analogous to what we describe here.

cAMP-cGMP cross-talk via PDEs
In addition to its direct effects on ion channels, cGMP can act via cross-talk with the cAMP pathway via its activation/inhibition of cellular PDEs (Zaccolo and Movsesian, 2007;Francis et al., 2010;Stangherlin et al., 2011). cGMP and cAMP compete with each other for the catalytic sites on the PDEs with varying affinities (Conti and Beavo, 2007;Francis et al., 2010), resulting in inhibition and/or activation of these enzymes. PDE2A and PDE3 have dual specificity for cAMP and cGMP (Conti and Beavo, 2007;Francis et al., 2010). cGMP activates PDE2A, leading to the breakdown of cAMP (and thus reduction in PKA-mediated phosphorylation of downstream targets including Ca 2ϩ channels), whereas it can inhibit PDE3A with the opposite result. Although a cell is likely to express several PDEs that regulate the efficacy of CNs, PDE2A and PDE3A have been localized in cardiac myocytes (Stangherlin and Zaccolo, 2012a;Maass et al., 2015;Zoccarato et al., 2015), where they are critically involved in cardiac function (Stangherlin and Zaccolo, 2012a). PDE2A has been reported recently in sympathetic stellate neurons with high activity in SHR neurons , whereas there is no direct evidence of PDE3 expression within sympathetic stellate neurons.
To test whether the cGMP dysregulation seen in the SHR was affecting cAMP signaling, we elevated cGMP levels pharmacologically and used FRET to record changes in cAMP levels. We investigated its functional relevance by assessing PKA activity. Application of 8b-cGMP led to a reduction of cAMP in SHR neurons, with no activation of PKA, suggesting that cAMP hydrolysis via cGMP stimulation of PDE2A is occurring. The cGMP-dependent reduction in the Ca 2ϩ channel current after 8b-cGMP is consistent with the notion of PDE2A dominance in diseased neurons affecting the efficacy of cGMP, because overexpressing PDE2A decreases cGMP and increases intracellular Ca 2ϩ transients and neurotransmission Seifert, 2015). Conversely, in the control neurons, even though 8b-cGMP significantly increased cAMP levels and PKA activity (suggesting the involvement of other PDEs inhibited by cGMP), this did not translate into a change in Ca 2ϩ channel current because activation of the inhibitory cGMP-PKG pathway most likely counterbalanced changes in the stimulatory cAMP-PKA pathway. However, slowing cGMP hydrolysis with PDE2A inhibition markedly increased the Ca 2ϩ current in normal neurons, probably as a consequence of cGMP acting to inhibit PDE3 leading to greater levels of cAMP-PKA activity ( Figure 5).

Evidence of altered PDE2A/PDE3A balance in sympathetic neurons
Mutations in PDE3A have been implicated in human hypertension (Houslay, 2015;Maass et al., 2015). Given the observed increase in cAMP after administration of 8b-cGMP in control neurons, we hypothesized that PDE3 may be functionally expressed within stellate neurons, as has been demonstrated in the superior cervical ganglions (Nunes et al., 2012). The selective PDE3 inhibitor milrinone led to similar increases in cAMP levels within both SHR and control neurons. Comparable results were also observed using the clinically approved PDE3 inhibitor cilostamide. The notion of differential PDE2/3 involvement in control and diseased neurons helps to clarify the effects that PDE2A inhibition has in the regulation of the Ca 2ϩ currents (Fig.  5). In control neurons, this leads to significantly increased currents, possibly due to the direct elevation of cAMP causing inhibition of cAMP hydrolysis, but also via the indirect elevation of cAMP mediated by cGMP inhibition of PDE3. This in turn may result in increased PKA-mediated phosphorylation of Ca 2ϩ channel currents. Conversely, in the SHR, the net result of PDE2A inhibition is a decrease in channel currents, possibly due to cGMP-mediated channel inhibition. This is consistent with the evidence that PDE2A has a higher affinity for hydrolysis of cGMP over cAMP (Xu et al., 2013;Li et al., 2015). Altered PDE2/3 balance could therefore be linked to the dysautonomia associated with hypertension (Fig. 8).

Limitations
CNs (Mongillo et al., 2006;Stangherlin et al., 2011;Stangherlin and Zaccolo, 2012b;Lomas and Zaccolo, 2014) and PDEs reside in distinct subcellular compartments (Lefkimmiatis and Zaccolo, 2014). Specifically, in the rat, there are three splice variants of PDE2A that are localized in the cytosol (PDE2A1) or membrane (PDE2A2 and 3;Bender and Beavo, 2006), although their exact roles remain to be established. Similarly, PKA/PKG regulation of the Ca 2ϩ channels also occurs in distinct signalosomes near the channels themselves, conferring site-specific regulation of Ca 2ϩ entry coupled to neurotransmission (Toda and Okamura, 2003;Garthwaite, 2016). Furthermore, the rate of PDE hydrolysis is critically dependent on the concentration of both cAMP and cGMP that is reported to be different between cell types (Francis et al., 2010;Zhao et al., 2016). Although the global cAMP responses after PDE2A and PDE3 inhibition were similar between control and diseased neurons, this does not in itself establish that localized control within specifically regulated microdomains is normal in the diseased neuron.