Reactive oxygen species alters the electrophysiological properties and raises [Ca2+]i in intracardiac ganglion neurons.

We have investigated the effects of the reactive oxygen species (ROS) donors hydrogen peroxide (H(2)O(2)) and tert-butyl hydroperoxide (t-BHP) on the intrinsic electrophysiological characteristics: ganglionic transmission and resting [Ca(2+)](i) in neonate and adult rat intracardiac ganglion (ICG) neurons. Intracellular recordings were made using sharp microelectrodes filled with either 0.5 M KCl or Oregon Green 488 BAPTA-1, allowing recording of electrical properties and measurement of [Ca(2+)](i). H(2)O(2) and t-BHP both hyperpolarized the resting membrane potential and reduced membrane resistance. In adult ICG neurons, the hyperpolarizing action of H(2)O(2) was reversed fully by Ba(2+) and partially by tetraethylammonium, muscarine, and linopirdine. H(2)O(2) and t-BHP reduced the action potential afterhyperpolarization (AHP) amplitude but had no impact on either overshoot or AHP duration. ROS donors evoked an increase in discharge adaptation to long depolarizing current pulses. H(2)O(2) blocked ganglionic transmission in most ICG neurons but did not alter nicotine-evoked depolarizations. By contrast, t-BHP had no significant action on ganglionic transmission. H(2)O(2) and t-BHP increased resting intracellular Ca(2+) levels to 1.6 ( +/- 0.6, n = 11, P < 0.01) and 1.6 ( +/- 0.3, n = 8, P < 0.001), respectively, of control value (1.0, approximately 60 nM). The ROS scavenger catalase prevented the actions of H(2)O(2), and this protection extended beyond the period of application. Superoxide dismutase partially shielded against the action of H(2)O(2), but this was limited to the period of application. These data demonstrate that ROS decreases the excitability and ganglionic transmission of ICG neurons, attenuating parasympathetic control of the heart.

parasympathetic regulation of the heart involves the convergence and integration of projections from the vagal motor nuclei within the intracardiac ganglia (ICG) (1). ICG represent the final common pathway through which the diverse, extrinsic neural signals to the heart are monitored, sending their projections to discrete regions of the heart.

Blockage of a coronary artery will result in inadequate blood flow downstream from the occlusion. Ischemia generates a multifaceted challenge, further complicated because chronic changes do not necessarily reflect augmented acute responses (25). Several factors underpin these changes, but reactive oxygen species (ROS) make a significant contribution. ROS are major players in the cascade of cellular injury that occurs both during ischemia (hypoxia) and reperfusion (reoxygenation).

Aerobic metabolism generates oxygen-derived oxidants—O2·−, H2O2, and·OH− as byproducts—which are increased in ischemia. Free radicals are generated by oxidative stress during ischemia coupled with a decreased availability of free radical scavengers. When this oxidant flux exceeds the capability of endogenous antioxidant mechanisms, tissue injury occurs. Restitution of blood flow (reperfusion) increases oxygen availability, which exacerbates damage due to increased production of ROS (6). Reperfusion in the clinical context occurs with thrombolytic agents or mechanically by angioplasty or emergency coronary bypass surgery, frequently resulting in impaired recovery, for example, development of arrhythmias (33).

Cardiac performance is regulated by both sympathetic and parasympathetic efferent drive. There is good evidence for interaction between the two in the ICG to regulate cardiac function (16, 17, 32). A recently published review describes the importance of a balanced cardiac sympathovagal drive for normal cardiac rhythm (22). Augmented sympathetic drive to the heart as part of the hemodynamic defense reaction to acute myocardial infarction and chronic heart failure has long been recognized. There has been much attention given to the detrimental, maladaptive effects of this aspect of the response as witnessed by the large cohort of research papers.

Impaired parasympathetic control of the heart is a powerful independent negatively prognostic predictor of arrhythmia and is a characteristic of myocardial infarction (43). An indication that decreased ganglionic transmission in ICG contributes to abnormal parasympathetic function in myocardial infarction comes from experimental models (11, 26).

The principal targets for the action of ischemia within the ICG are 1) synaptic transmission and 2) the evoked discharge pattern of the post-ganglionic neuron.

Experimentally, different ROS-generating and/or endogenous ROS-identifying systems have been used to examine ROS-induced changes in electrophysiological properties and synaptic responses. ROS donors include H2O2 and tert butyl hydroperoxide (t-BHP), a substrate of glutathione peroxidase. ROS-induced effects have been confirmed by examining the effects of free-radical scavengers of O2·−, H2O2, and·OH− such as catalase and superoxide dismutase (SOD).

The proximity of the intracardiac ganglia to the coronary blood supply makes them susceptible to the effects of ROS (2). H2O2 administered to the blood supply of canine ICG ganglia in situ attenuated neuronal firing (46). ROS produced by the myocardium during ischemia-reperfusion have been shown to alter the firing properties of cardiac sensory neurites associated with afferent axons in vagal and sympathetic nerves (49).

Ion channels and transporters are susceptible to the action of ROS. For example, voltage-dependent Na+, K+, and Ca2+ channels, Ca2+-activated K+ channels, and KATP channels have all been identified as targets for ROS (23, 27). A recent paper reported that ROS donors (H2O2 and t-BHP) reduced the voltage operated calcium current but increased the amplitude of the delayed rectifier K+ current in dissociated ICG neurons (55).

Calcium homeostasis, for example, the ion transporters regulating intracellular Ca2+ levels, is disrupted by ROS (27). Increases in Ca2+ initiates inappropriate activation of several enzyme systems e.g., nitric oxide synthase and phospholipase A2. Overactivation of these enzymes results in the breakdown of proteins and phospholipids and initiates several cascades that damage cells (28).

Here, we report the action of the ROS donors H2O2 and t-BHP, simulating one component of ischemia upon the intrinsic passive and active properties, ganglionic transmission, and [Ca2+]i in the neurons of the ICG, which regulate the sinoatrial node.

H2O2 and t-BHP generate distinct ROS species. H2O2 produces the free radicals superoxide O2− and hydroxyl OH. (27). By contrast, the oxidant t-BHP is rather stable in solution (50), but it triggers the generation of free radical intermediates peroxyl and alkoxyl radicals, which can cross cellular membranes and evoke the production of the hydroxyl radicals (24).

We tested the hypotheses that 1) ROS attenuates synaptic responses in ICG neurons, 2) ROS increases [Ca2+]i, and 3) the distinct complement of ion channels and receptor ion channels expressed by neonatal and adult ICG neurons modulates their sensitivity to ROS.

Some of the actions of H2O2 and t-BHP upon the intrinsic and active properties of ICG neurons have been published in an abbreviated form previously (14, 55). The present report has enhanced these studies by investigating the action of ROS donors and scavengers on ganglionic transmission and Ca2+ homeostasis, examining both adult and neonatal ICG neurons. In addition, it has extended our understanding of the likely targets for the action of ROS.

To our knowledge, there are presently no reports on the action of ROS-donors on ganglionic transmission in intracardiac ganglion neurons.

MATERIALS AND METHODS

Preparation.

The whole mount ICG preparation has been described previously (38). Briefly, Wistar rats (Harlan UK, Oxon, UK) were used at two stages of postnatal development: neonates (P2-P9) and young, nonpregnant, female adult Wistar rats (≥6 wk, 125–220 g). The University of Dundee is a designated scientific establishment (certificate of designation no. 60/2602) under the Animals (Scientific Procedures) Act 1986 (“the Act”). Rats were obtained from a designated supplier in the UK and were housed and cared for according to the Home Office Guidelines on the Operation of the Act. Animals were killed by concussion and cervical dislocation, as authorized in Schedule 1 to the act. A whole mount preparation comprising the right atrial ganglion plexus and underlying myocardium was pinned out in a recording chamber (∼1.0 ml volume) lined with Sylgard 184 silicone elastomer (Dow Corning, Barry, UK) and superfused with bicarbonate buffered physiological salt solution (PSS) at ∼2 ml/min (Gilson Minipuls 3; Gilson, Bedford, UK). The temperature of the superfusing solution was controlled by a Peltier heating device (Medical Systems PDMI-2 micro incubator; Medical Systems Corp., Greenvale, NY) to 36°C, monitored by an independent thermistor probe in the recording chamber. The tissue was left to resuscitate in these conditions for ∼30 min before commencing recording. ICG neurons were visualized using differential interference contrast (DIC) optics on a fixed stage microscope. Recordings were normally made from the sino-atrial ganglion, the largest located at the junction of the right superior vena cava and right atrium (41).

Electrophysiological recording, data acquisition and analysis.

Intracellular recordings from postganglionic ICG somata were made using sharp microelectrodes pulled from thin-walled borosilicate glass (GC120F; Harvard Apparatus, Kent, UK) with resistances of ∼120 MΩ when filled with 0.5 M KCl. Membrane voltage responses were recorded with a conventional bridge amplifier (Axoclamp 2A; Axon Instruments, Foster City, CA). Voltage signals were filtered at 20 kHz (Frequency Devices 902; Frequency Devices, Ottawa, IL), digitized at 50 kHz and transferred to a dual-core Pentium computer using an analog-to-digital converter [Micro 1401 Mk II interface; Cambridge Electronic Design (CED), Cambridge, UK] and Spike 2, version 6 software (CED).

Two types of current clamp protocol were routinely performed. In the first, brief intracellular depolarizing currents (≤3 ms in duration) were used to directly evoke single somatic action potentials. Action potential parameters measured were overshoot, afterhyperpolarization (AHP) amplitude and duration to 50% recovery (AHP50), using a Spike 2 script. Long (500 ms) hyperpolarizing and depolarizing pulses were used to measure time-constant and time-dependent rectification, and evoked discharge characteristics, respectively. Time constant (τ) was calculated from the voltage response to small hyperpolarizing, long-current pulses (≤ − 0.1 nA) using Spike 2 software. Membrane resistance (Rm) was calculated from τ = Rm × Cm, where Cm is the specific membrane capacitance (assumed to be 1 μF·cm−2). Discharge activity was classified as being phasic, multiple adapting, or tonic upon application of a long depolarizing current pulse approximately twice threshold intensity.

Branches of the vagus and interganglionic nerve trunks were stimulated using a glass suction electrode connected to a constant voltage isolated stimulator (Digitimer DS2; Digitimer, Welwyn Garden City, UK). Nerve trunks were stimulated using stimulus pulses of 0.02 to 0.2 ms width and 5–50 V amplitude. Nicotine were focally applied using a pressure-ejection device (∼150 kPa; Picospritzer II, General Valve, Fairfield, NJ), and the pressure ejection pipette was positioned <50 μm from the neuronal soma to maximize the response to agonist application.

Solutions and pharmacological agents.

PSS contained (in mM): 118 NaCl, 25 NaHCO3, 1.13 NaH2PO4, 4.7 KCl, 1.8 CaCl2, 1.3 MgCl2, 11.1 glucose and was gassed with 95% O2-5% CO2 to pH 7.4 (45). For calcium-free PSS no CaCl2 was added, MgCl2 was increased to 3.9 mM and contained 0.5 EGTA mM but was otherwise similar to normal PSS (31). Pharmacological agents were prepared and dissolved immediately before application in PSS at the concentrations stated. All reagents were of analytical grade.

Intracellular calcium measurements.

Resting intracellular calcium levels and dynamics were measured using a back-illuminated electron multiplying gain charge-coupled device camera (DU-860E, 128×128 pixels; Andor Technology, Belfast, Northern Ireland) to measure fluorescence of the nonratiometric membrane impermeable calcium indicator, Oregon Green 488 BAPTA-1 (OGB-1; Molecular Probes, Invitrogen, Carlsbad, CA). OGB-1 is a high-affinity Ca2+ indicator (Kd ∼170 nM), an advantage for detecting small changes in Ca2+ near resting values. Imaging was carried out using epifluorescence optics (Leica filter cube L5). Intracellular recordings were made using sharp glass microelectrodes whose tip was filled with 0.5 mM OGB-1 (dissolved in 200 mM K+ acetate) and backfilled with 3 M K+ acetate (54), allowing simultaneous recording of electrophysiological properties and measurement of [Ca2+]i. Images were captured and analyzed using National Instruments M-Series interface card (National Instruments, Austin, TX) and Winfluor version 3 software (authored by Dr. John Dempster, University of Strathclyde, UK) running on a dual-core Pentium computer. OGB-1 was injected iontophoretically with hyperpolarizing current pulses (−0.1 nA, 500 ms, 1 Hz) until the fluorescence reached a steady state, ∼10 min. The dye quickly spread out evenly throughout the cytoplasm. The morphology of the somata of rat ICG neurons is typically placentiform, having few, if any, short dendrites. Therefore, the soma presents an excellent site for recording changes in [Ca2+]i. Maximum fluorescence (representative of maximum calcium levels, fmax) was measured in response to rapid trains of directly evoked action potentials at various frequencies (5–100 Hz). Dye saturation was tested by comparing the fluorescence transients produced by trains at different frequencies (data not shown). Minimum fluorescence (fmin) was measured by switching the superfusing normal PSS to calcium-free PSS, resulting in a decrease in baseline fluorescence (fo). Signals obtained due to changes in [Ca2+]i were expressed as the ratio of fluorescence changes over baseline fluorescence, f − fo/fo.

Resting intracellular free [Ca2+] was calculated using the equation: where R and Kd (in nM) are the dynamic range and dissociation constant of OGB-1 whose values are 5 and 206, respectively. (30). Throughout the experiment, excitation parameters (light intensity, scan duration) were adjusted to minimize the photobleaching of OGB-1.

Statistics.

Data are presented as the means ± SD and were compared using one-way repeated ANOVA (Holm-Sidak) and paired t-tests (SigmaStat 3.1; Systat Software, Chicago, IL).

RESULTS

General properties of ICG neurons.

To be included in this study, ICG neurons had a resting membrane potential (Em) ≥ −40 mV and overshooting somatic action potentials elicited by short depolarizing current pulses (2–3 ms) under control conditions. Recordings were stable for at least 10 min before taking readings and altering the superfusing PSS solution. Many neurons in both neonate and adult ICG displayed excitatory postsynaptic potential (EPSP) activity (5/6 neonates and 21/64 adults), and occasionally spontaneous action potentials (APs 3/6 neonates and 6/64 adults).

The action of H2O2 and t-BHP on the passive membrane properties of the postganglionic neuron.

The effects of exogenously applied ROS-generating systems hydrogen peroxide (H2O2) and t-BHP on the electrophysiological properties in ICG neurons have been investigated over an exposure time period of 20 min. Data were normally taken at two time points in control conditions (−10, 0 min), and following 20 min superfusion of the ROS donor.

The membrane potential (Em) response of an adult ICG neuron to superfusion of H2O2 (1 mM) and subsequent coapplication of Ba2+ is displayed in Fig. 1. The time dependence of action of H2O2 (1 mM) on the resting Em in adult ICG neurons is shown in Fig. 1. For all neurons, an early response (depolarizing or absent) was followed by a slowly developing hyperpolarization, which increased with exposure time to a steady state (see Fig. 1). This negative change in Em evoked by H2O2 was not reversed upon washout (see Fig. 1). This lack of reversibility also applies to t-BHP (data not shown). This observation is, however, limited to short periods of washout; dislodgement of the microelectrode caused by contraction of the underlying atrial musculature frequently curtailed recordings.

Fig. 1.

Effect of H2O2 and tert-butyl hydroperoxide (t-BHP) on the resting membrane potential (Em) in adult rat ICG neurons in situ. A: membrane potential responses of an intracardiac ganglion (ICG) neuron to superfusion of H2O2 (1 mM) and the subsequent coapplication of Ba2+. Sections of the recording show the early, transient, depolarization, and ensuing hyperpolarization. Spontaneous action potential discharge was recorded during the initial depolarizing response and with application of Ba2+ (indicated by arrows). Ba2+ (1 mM) reverses the membrane hyperpolarization produced by H2O2. B: time dependence of action of H2O2, the mean response from 15 neurons is presented. C: hyperpolarization produced by H2O2 is not reversed upon washout. Data are presented here for five neurons for which recordings were stable for >10 min following reverting to control conditions after H2O2. D: early membrane potential response and change following 20 min superfusion of H2O2 is presented, both as individual and as mean ± SD, for neurons displaying a transient depolarization (i) and neurons exhibiting no depolarizing response (ii). E: concentration-response relationship for changes in resting Em produced by H2O2, (data presented as Em changes relative to control rather than absolute Em). F: t-BHP also produces a membrane potential hyperpolarization. The membrane potential response to 20 min superfusion t-BHP (1 mM) is shown; data are presented both as individual responses and as the mean ± SD. **P < 0.01; ***P < 0.001; ns, not significant.

Considering the two distinct responses recorded immediately upon exposure of the neuron to H2O2, in one group of neurons (25/39), superfusion of H2O2 produced a significant transient, small, depolarization, sometimes sufficient to evoke action potential discharge (from −49.2 ± 4.9 mV control to −45.5 ± 5.2 mV H2O2; P < 0.001) (see Fig. 1, A and D). In the remainder (14/39), this early depolarizing response was absent. The steady-state hyperpolarization, taken at 20-min exposure to H2O2, was the same for neurons with an early depolarizing response and those in which it was absent [from −49.2 ± 4.9 mV control to −63.8 ± 8.3 mV H2O2 (P < 0.001; n = 25) for those neurons displaying an initial transient depolarization; from −48.6 ± 5.1 mV control to −69.4 ± 8.6 mV H2O2 (P < 0.001; n = 14) for neurons showing only a hyperpolarizing response) (see Fig. 1).

The effects of varying concentrations of H2O2 (0.1–1 mM) on adult ICG neurons were investigated to characterize the concentration dependence of its action. H2O2 at concentrations of 0.1 and 0.2 mM (n = 5) produced no significant Em changes. Application of 0.5 and 1 mM H2O2 changed resting Em (from −48.4 ± 4.7 mV in control to −55.5 ± 7.3 mV in 0.5 mM H2O2; and from −48.4 ± 4.7 mV in control to −68.4 ± 8.9 mV in 1 mM H2O2; P < 0.01; n = 5, 20 min) (see Fig. 1). The concentration of H2O2 used elsewhere in the manuscript refers to 1 mM.

The change in Em evoked by H2O2 (1 mM, 20 min) in neonatal neurons was not different from that recorded from adult ICG neurons (−16.8 ± 7.8 mV, n = 39 in adults and −12.0 mV ± 7.7, n = 6 in neonates) (see Supplemental Table 1 in the online version of this article).

Superfusion of t-BHP (1 mM) also produced a hyperpolarization of resting Em (see Fig. 1 and Supplemental Table 1). There was no transient depolarizing response with superfusion of t-BHP. The magnitude of the hyperpolarizing shift in resting Em produced by H2O2 in adult ICG neurons was significantly greater than that for t-BHP (−16.8 ± 7.8 mV in H2O2, n = 39 vs. −7.2 mV ± 5.7 in t-BHP, n = 21; P < 0.01). The action of t-BHP on neonatal neurons was not examined.

The hyperpolarization of resting Em caused by both H2O2 and t-BHP was associated with a decrease in membrane resistance (Rm, measured at ≤ −0.1 nA) (see Supplemental Table 1).

H2O2 mediated membrane hyperpolarization involves K+ channel activation.

The ionic conductances underlying the H2O2 induced Em hyperpolarization were investigated by exposing adult ICG neurons to Ba2+, a wide spectrum K+ channel blocker. Ba2+ (1 mM) reversed the H2O2 induced hyperpolarization; indeed, the Em settled slightly positive to control values (−49.8 ± 5.3 mV in control, −68.6 ± 7.6 mV in H2O2 to −46.8 ± 6.5 mV in H2O2 + Ba2+; n = 11; P < 0.001) (see Fig. 2), confirming K+ channel involvement. The depolarization caused by Ba2+ was associated with an increase in Rm (8.1 ± 2.1 kΩ·cm2 control, 5.0 ± 3.1 kΩ·cm2 in H2O2 and 13.0 ± 4.5 kΩ·cm2 in H2O2 + Ba2+; P < 0.05; n = 6).

Fig. 2.

The actions of K+ channel blockers on H2O2 induced Em changes and excitability in adult rat ICG neurons. A and B: resting Em recorded in control, following H2O2 superfusion (20 min) and following coapplication of the wide spectrum K+ channel blocker Ba2+ (1 mM), or the voltage-gated K+ channel blocker tetraethylammonium (TEA) (10 mM), C: membrane potential response to brief, depolarizing current pulses (i) control (+0.4 nA), evokes a somatic potential (ii) H2O2, the negative shift in Em and reduced membrane resistance results in an increased (+ 0.8 nA) current intensity evoking only a subthreshold voltage response (iii) +0.4 nA, application of Ba2+ (in the continued presence of H2O2) reverses the H2O2 depolarization and increases membrane resistance restoring excitability. D: SK channel blocker apamin (100 nM) has no impact on the H2O2 negative shift in Em E-G the actions of the muscarinic ACh agonst muscarine (20 μM) and M-channel blockers on the H2O2-evoked hyperpolarization of Em: E: muscarine (20 μM). F: oxotremorine (10 μM). G: linopirdine (10 μM). *P < 0.05, **P < 0.01, ***P < 0.001. ns, not significant.

The involvement of specific K+ channel blockers was assayed using selective blockers, applied when the hyperpolarizing shift in Em was fully developed (20 min). Tetraethylammonium (TEA; 10 mM), an inhibitor of voltage and Ca2+-activated K+ channels, to an extent reversed the H2O2-induced hyperpolarization (−48.1 ± 4.1 mV in control, −64.5 ± 8.6 mV in H2O2, and −56.5 ± 11.6 mV in H2O2+TEA, n = 5, P < 0.05) (see Fig. 2).

Somatic action potentials were normally produced in response to injection of brief depolarizing current pulses in control conditions; however, strong currents evoked only a subthreshold voltage response in H2O2 (see Fig. 2). The diminished excitability presumably results from the negative shift in Em and reduced membrane resistance. Application of Ba2+ (in the continued presence of H2O2) reverses the H2O2 depolarization and restores excitability (see Fig. 2).

Apamin (the small calcium activated K+ channel SKCa blocker, 100 nM) did not reverse the action of H2O2 on Em (−46.7 ± 4.5 mV control, −67.8 ± 8.6 mV H2O2, and −63.3 ± 10.8 mV H2O2+apamin; n = 7) (see Fig. 2).

The participation of the muscarine-sensitive K+ channel was tested using muscarine (M-current inhibitor, 20 μM). This agent partially abrogated the H2O2-induced hyperpolarization (−48.7 ± 5.9 mV control, −66.6 ± 5.2 mV H2O2 and −57.2 ± 8.0 mV in H2O2 + muscarine, n = 8, P < 0.01) (see Fig. 2). In line with decreasing a K+ conductance, muscarine increased Rm (7.4 ± 2.4 kΩ·cm2 control, 4.6 ± 1.8 kΩ·cm2 in H2O2 to 7.7 ± 3.6 kΩ·cm2 in H2O2 + muscarine; P < 0.05; n = 6). The M-current was further investigated in the presence of Cs+ (3 mM), TTX (300 nM), and 4-AP (1 mM) used to isolate the M-current (9). The H2O2-evoked hyperpolarization shift while superfusing this mixture was not different from the membrane potential change produced by H2O2 alone (−16.1 mV ± 5.3, n = 5 and −16.8 ± 7.8 mV, n = 39, respectively).

Considering M-channel blocker analogs, oxotremorine-M (10 μM) and XE-991 (50 μM, n = 3, data not presented) did not show any effect (see Fig. 2). Linopirdine, an open channel blocker (10 μM), partially reversed the H2O2-induced hyperpolarization (−49.2 ± 3.2 mV control, −62.1 ± 6.7 mV H2O2, and −58.1 ± 5.0 mV H2O2+linopirdine; n = 4; P < 0.05) (see Fig. 2).

In neonatal neurons the H2O2 induced hyperpolarization was completely reversed by TEA, 10 mM (−50.0 ± 5.4 mV control, −59.2 ± 3.3 mV H2O2, −49.4 ± 9.5 mV H2O2+TEA; n = 5; P < 0.05).

Time-dependent rectification.

Application of hyperpolarizing current pulses can induce time-dependent rectification (TDR), held as the signature of the H-current, in ICG neurons (35). Such behavior was either blunted or absent in H2O2 in adult ICG neurons (see Fig. 3). Similar behavior was observed upon superfusion of t-BHP. The extent of TDR was quantified by measuring steady-state voltage response to a hyperpolarizing current pulse to approximately −90 ± 10 mV and expressing this as a percentage of the peak, with time, membrane potential excursion. The values were 0.91 ± 0.03% and 0.98 ± 0.02% (n = 19, P < 0.001) in control and H2O2, respectively; and 0.93 ± 0.06% and 0.97 ± 0.03% (n = 7, P < 0.05) in control and t-BHP, respectively. Neonatal ICG neurons showed less TDR in control PSS, 0.96 ± 0.03 (n = 5), in agreement with a previous report (38). There was no significant action of H2O2 on TDR in neonatal ICG.

Fig. 3.

Membrane potential response to depolarizing and hyperpolarizing current pulses and evoked discharge characteristics. A: voltage responses (i and ii) obtained in response to depolarizing and hyperpolarizing current pulses (+0.2, −0.1, to −0.5 nA) are shown for control conditions and after 20-min exposure to H2O2 (1 mM). A: current-voltage (I-V) relations (iii and iv) for each condition plotted for the peak membrane potential excursion (○) and the steady state response measured at the end of the current step (●). B: examples of the action of H2O2 (1 mM) on the discharge activity of two neurons classified as tonic (i) and phasic (ii) firing in response to a 500-ms twice-threshold depolarizing current pulse, upper traces control, lower traces taken after 20-min superfusion of H2O2. C: block histograms depicting changes in evoked firing patterns produced by H2O2 (1 mM) (i) and t-BHP (1mM) (ii).

Active properties.

ROS donor-induced hyperpolarization was associated with a decrease in excitability; this was manifest as a switch from predominantly phasic or multiple adapting discharge to phasic or unresponsive (see Fig. 3).

Several membrane currents that are involved in the regulation of repetitive activity and adaptation have been characterized in rat ICG neurons. The inventory includes Ca2+-dependent K+ currents (IK,Ca), the transient outward K+ current (IA), the muscarine-sensitive K+ current (IM), and the hyperpolarization-activated, nonspecific cation current (Ih). These currents, in the main, are activated in the voltage range between the peak AHP following the action potential and action potential threshold and can affect the general level of excitability. IM and Ih also contribute to the resting membrane potential in these neurons. Ba2+ and TEA reversed the H2O2-induced switch in evoked firing patterns (3/3 and 2/2 neurons, respectively).

Somatic action potential (AP) parameters measured were overshoot (OS), rate of rise (drise/dt), rate of fall (dfall/dt) and the AHP following the action potential, characterized by its depth below resting Em (AHP) and time to 50% recovery AHP50 (15). Adult ICG neurons had APs with large AHP amplitudes (18.5 ± 6.1 mV, n = 40) and a wide range of AHP50 durations (7.5–50.5 ms), in accordance with previously reported values (13, 38). Application of H2O2 reduced the AHP in adult and neonatal ICG neurons. Similarly, t-BHP reduced the AHP in adult ICG neurons (see Supplemental Table 1). ROS donors had no impact on OS or AHP50 in either adult or neonatal ICG neurons. Considering adult ICG neurons, the maximum rate of rise (max drise/dt) and fall (max dfall/dt) of the action potential were decreased by t-BHP (max drise/dt rise from 162.7 ± 58.6 V/s control to 118.5 ± 57.0 V/s, n = 6, P < 0.05 and max dfall/dt from 61.1 ± 21.9 V/s control to 57.0 ± 22.6 V/s, n = 6, P < 0.01). This parameter was not available for H2O2 simply because of the Em hyperpolarization and decreased Rm-induced loss of excitability (28/39 neurons).

The action of H2O2 on the nicotinic ACh receptor was assayed by focal application of nicotine. Short pulses of nicotine (100 μM, 20 ms) evoked transient depolarizing responses and action potential discharge (see Fig. 4). The membrane potential response to nicotine was not altered by H2O2 (see Fig. 4). Whereas the peak, with time, depolarizing response evoked by nicotine was increased, the Em value attained was unchanged: −30.6 ± 9.3 mV in control PSS and −32.0 ± 9.0 mV, n = 4 in H2O2 (20 min). The action potential discharge associated with the fast depolarizing response to nicotine was, however, normally absent in H2O2 (see Fig. 4).

Fig. 4.

Effect of reactive oxygen species (ROS) on somatic action potentials and nicotine evoked Em responses in ICG neurons. A: examples of the action of ROS on somatic action potentials (APs) recorded from adult (i and ii) and neonatal (iii) intracardiac neurons. Somatic AP traces were evoked by brief 2-ms depolarizing current pulses, control APs are shown in black and for H2O2 and (i and iii) and t-BHP (ii) in gray. B: focal application of nicotine, (100 μM, 20 ms) indicated by arrowhead evoked transient depolarizing responses and action potential discharge (inset). The nicotine-evoked depolarization was unaffected by H2O2. Responses taken in control (i), 10 (ii), and 20 min (iii) following application of H2O2 are presented.

Ganglionic transmission.

All ICG neurons included in this study received a strong synaptic input, i.e., they received a suprathreshold EPSP, which evoked an action potential in response to each supramaximal stimulus applied to the preganglionic nerve trunk at 0.2 Hz (13). Ganglionic transmission was investigated both for low-frequency stimuli (0.2 and 0.5 Hz) and multiple trains of stimuli applied at 5–100 Hz. Transmission was irreversibly blocked by H2O2 in almost all neurons (14/15). This failure of transmission was progressive, i.e., secure APs were first reduced to subthreshold EPSPs before complete block of the postganglionic response. (see Fig. 5, A and B). The absence of a nerve-evoked response in H2O2 could be the result of axonal conduction block or lack of action of ACh on the postganglionic neuron. The lack of action of H2O2 on the postsynaptic response to nicotine (presented in Fig. 4) indicates that the latter can be dismissed. Considering conduction block, there was no change in the latency of evoked responses to single stimuli before blunting or block of ganglionic transmission (2.8 ± 3.0 ms in control and 1.7 ± 0.9 ms in H2O2, 20 min, n = 5). In addition, H2O2 had no effect on antidromic conduction (data not shown).

Fig. 5.

The actions of H2O2 and t-BHP on ganglionic transmission in adult ICG neurons. A: progressive block of ganglionic transmission by H2O2 (1 mM). Waterfall display of nerve-evoked responses to single stimuli (s, applied at 0.2 Hz) taken at 60-s intervals. H2O2 was applied at the time indicated by an arrow. Ganglionic transmission failed ∼12 min after application of H2O2. B: action of ROS on ganglionic transmission. The effect of H2O2 (i) and t-BHP (ii) on indirect nerve-evoked responses (stimulation applied at time “s”), control shown in black, ROS (20 min) in gray. C: attenuation of postsynaptic events evoked by trains of nerve stimuli with H2O2 from an adult ICG neuron. Waterfall display of 20 stimuli at 10 Hz applied at time “s” is shown. Responses to a train of such stimuli delivered to the preganglionic nerve trunk in control conditions (i) and following superfusion of H2O2 (20 min, ii) are presented. D: effects of ROS on synaptic transmission were investigated by applying trains of stimuli (5–100 Hz) to the preganglionic nerve trunk, and the ratio of the number of successful somatic action potentials to the number of stimuli gives an index of the frequency dependency of ganglionic transmission. This analysis is presented for adult neurons in control conditions and following application of H2O2 (i) and t-BHP (ii). ***P < 0.001.

The ability of the postganglionic neuron to follow the activity of preganglionic stimuli was investigated by applying trains of 20 stimuli (twice threshold voltage) at different frequencies up to 100 Hz. The ratio of the number of successful action potentials to the number of stimuli was used to provide an index of the frequency dependence of ganglionic transmission. The ability of the postganglionic neuron to faithfully follow preganglionic nerve stimulation decreases at high frequencies. Ganglionic transmission was blocked over the range of frequencies examined (5–100 Hz) by H2O2 (see Fig. 5, C and D).

In contrast, t-BHP had no action on ganglionic transmission. Congruent with its action on APs evoked by direct stimulation, it decreased the synaptic AP rate of rise from 159 ± 79 V/s control to 117 ± 53 V/s (n = 6, P < 0.05) and increased AP duration (measured at 0 mV) from 0.7 ± 0.2 ms to 1.1 ± 0.3 ms (n = 6, P < 0.05; see Fig. 5). The action of t-BHP on the frequency dependence of ganglionic transmission is shown in Fig. 5.

The action of H2O2 following pretreatment with superoxide dismutase and catalase (adult ICG neurons).

Antioxidant enzymes such as SOD and catalase provide protection against ROS-mediated ischemia-reperfusion injury (56). To test the action of these agents, ICG were preexposed to either 100 U/ml SOD or 100 U/ml catalase for 10 min followed by the coapplication of H2O2 and antioxidant for 20 min.

Application of SOD in itself had no action on the electrical properties and ganglionic transmission of ICG neurons (see Supplemental Table 2 in the online version of this article). Catalase caused a slight, but significant, hyperpolarization of the resting Em (from −47.3 ± 4.3 mV, control to −48.7 ± 5.0 mV, catalase; P < 0.05; n = 8) and improved the security of ganglionic transmission at high frequencies: 100 Hz (0.82 ± 0.30 control, 0.86 ± 0.27 catalase, n = 5, P < 0.05) (Fig. 6, A and C).

Fig. 6.

Catalase abrogates and superoxide dismutase (SOD) obtunds the membrane potential hyperpolarizing and ganglionic transmission blocking actions of H2O2. A and B: effects of the ROS scavengers catalase and SOD (100 units/ml), ROS scavengers combined with H2O2 (1 mM), and H2O2 alone on resting Em of adult ICG neurons. By contrast with SOD, the shielding action of catalase extends beyond the time period of its application. The gray rectangle in and A and B depicts ± 2 × SD of the mean of control values. C and D: effects of catalase or SOD, combined catalase, or SOD and H2O2, and H2O2 alone on ganglionic transmission in the adult ICG. Trains of 20 stimuli (5–100 Hz) were applied to the preganglionic nerve trunk. The ratio of the number of successful somatic action potentials to the number of stimuli gives an index of the frequency dependency of ganglionic transmission. *** P < 0.001.

The action of coapplication of SOD and H2O2 was not consistent. In 2/5 neurons, superfusion of SOD and SOD plus H2O2 produced no change in postganglionic neuron electrical properties and ganglionic transmission. Subsequently, H2O2 alone produced a hyperpolarization of Em. In the remaining three neurons, SOD did not protect against the action of H2O2 (see Fig. 6). The action of H2O2 in both these instances was associated with reduced TDR, a shift in evoked firing pattern and ganglionic transmission blockade.

Superfusion of catalase and H2O2 showed no significant differences in postganglionic membrane properties and ganglionic transmission compared with that of catalase in all of the neurons studied (Supplemental Table 2). This protective action of catalase extended beyond the period of its application. Application of H2O2 following coapplication with catalase did eventually evoke its expected effects (Em hyperpolarization, reduction in τ and Rm, blunted TDR, and altered evoked firing patterns) but only after 30-min to 1-hr exposure.

Intracellular calcium.

ROS donors increased the resting [Ca2+]i in adult ICG neurons. The time dependence of action of H2O2 on resting [Ca2+]i is shown for an individual neuron in Fig. 7. Both H2O2 and t-BHP markedly increased resting [Ca2+]i to 1.6 ± 0.6 (n = 11, P < 0.01) and 1.6 ± 0.3 (n = 8, P < 0.001), respectively, of control values (Fig. 7, A and C). However, this increase was delayed from the onset, occurring after ∼15-min exposure (see Fig. 7). Because of the relatively prolonged duration of these procedures, the extent of photobleaching and/or transport of OGB-1 from the soma of the ICG neuron was a matter of concern. To gauge this, fluorescence was measured in 31 OGB-1-loaded neurons over a period of 50 min. A linear fit regression provided the best description of the time course of bleaching (y = 0.977 −0.130, r2 = 0.98). This best fit ± 2 × SD is plotted on the graphs for the actions of H2O2 and t-BHP on [Ca2+]i in Fig. 7.

Fig. 7.

The increase in intracellular calcium in adult rat ICG neurons by ROS donors and the shielding action of catalase. A: epifluorescence images (i) of an OGB-1-filled adult rat ICG neuron taken in control conditions and following 20 min of superfusion of H2O2 (1 mM). The microelectrode is shown at the right side of the neuron. A: [Ca2+]i changes (ii) with time in H2O2 expressed as the ratio of fluorescence changes over baseline fluorescence, (f − fo/fo). B: epifluorescence images (i) of an OGB-1 filled adult rat ICG neuron taken in control conditions and following 10 min of superfusion catalase and 20 min coapplication of catalase and H2O2 (1 mM). B: [Ca2+]i changes (ii) with time in H2O2 expressed as the ratio of fluorescence changes over baseline fluorescence, (f − fo/fo). C: actions of ROS on fluorescence intensity of neurons loaded with OGB-1. For both H2O2 (i) and t-BHP (ii), data are presented both as individual results and as the ± SD of the mean (taken following 20-min exposure to the ROS). The dotted line and gray triangle depict the mean ± 2 × SD of the mean of time-dependent photobleaching of OGB-1. *** P < 0.001.

The addition of catalase had no significant effect on resting [Ca2+]i levels nor did H2O2 coapplied with catalase (f − fo/fo 1.0 ± 0 control, 1.0 ± 0.2 catalase, 0.8 ± 0.2 catalase + H2O2, 0.6 ± 0.2 H2O2, n = 5). Akin to its guarding against the action of H2O2 on the electrical properties of ICG neurons, the protective effect of catalase outlived the timeframe of its application (see Supplemental Table 2 and Fig. 7).

Resting [Ca2+]i was estimated to be 62 ± 24.3 nM, n = 30 (12), which is in agreement with the sparse data available on [Ca2+]i in dissociated autonomic ganglion neurons (4, 53).

DISCUSSION

The key observations made in this study are that ganglionic transmission is blocked by H2O2. The sensitivity of the nicotinic ACh receptors on the postganglionic neuron was unaffected indicating that ganglionic block is due to a presynaptic action of H2O2. Considering the postganglionic neuron, both ROS generating agents evoked a membrane potential hyperpolarization and associated decrease in membrane resistance resulting in a decrease in its excitability. This action was fully reversed by Ba2+ (a wide-spectrum K+ channel blocker). Both ROS-generating agents increased intracellular [Ca2+] in ICG neurons. H2O2 and t-BHP switched the evoked discharge characteristics of ICG neurons from phasic/multiple adapting to inexcitable/phasic. ROS scavengers had shielding actions against the effects of ROS donors on Em, ganglionic transmission and Ca2+ homeostasis. Together, the actions of ROS on ganglionic transmission and excitability of the postganglionic neuron will have a parasympatholytic effect.

We used 1 mM concentrations of H2O2 and t-BHP in these studies. Are these concentrations physiologically relevant and what are the likely targets of ROS action? Considering neuronal function, ROS can attack ion channels and transporters directly, or indirectly by causing lipid peroxidation (6, 27) and affecting associated signaling molecules (23). In contrast to some other changes associated with ischemia, for example, increased [K+]o, ROS can target both the plasma and intracellular membranes, e.g., mitochondria as well intracellular signaling mechanisms. A wide range of concentrations of exogenously applied ROS donors have been used to simulate the accumulation of oxygen-derived free radicals occurring during ischemia and reperfusion. A review of the action of H2O2 indicates that exogenous application of concentrations up to 1 mM will replicate endogenous release and are directly relevant to cardiovascular studies (42). In the time frame of our studies, evidence indicates that lipid peroxidation is unlikely to be a major player. Only mild nonspecific lipid peroxidation of membrane lipids by H2O2 takes place in 20 min in isolated nerve terminals (48). Interestingly, H2O2 decreased ATP levels to ∼60% within 30 min (whereas t-BHP requires 2 h to achieve the same reduction) (36).

A wide range of levels of ROS scavengers have been used in cell and tissue studies. We elected to use catalase and SOD at 100 U/ml, which is approximately the midrange of previous comparable studies (3, 7, 10, 34, 55).

Investigation of ganglionic transmission in the present study was limited to those postganglionic neurons receiving a secure, suprathreshold, input from the vagus, presumed to form efferent outflow (13). These neurons would be classified as principal cardiac neurons according to the cellular morphological scheme of Cheng and Powley (8). H2O2 blocked ganglionic transmission at all frequencies. This could be the result of suppression of exocytosis at the preganglionic nerve terminals or block of the postsynaptic response. The action of nicotine on the postganglionic membrane was unaffected by H2O2, indicating that the nicotinic ACh receptor-channel complex is resistant to the action of ROS. Previous studies have shown that H2O2 at high concentrations (mM) depresses synaptic transmission both in the CNS (7) and at the neuromuscular junction (20). The presynaptic action of H2O2 in frog neuromuscular junction was demonstrated to be due to its action on the neurotransmitter release protein SNAP25 (19).

In sympathetic ganglion neurons, increases in intracellular levels of ROS also depressed synaptic transmission, but by a rundown of ACh-evoked currents (5). The biosynthesis and degradation of catecholamines in addition to their autooxidation make sympathetic postganglionic neurons particularly susceptible to oxidative damage (5, 21). This difference in neurochemistry may well underpin the differences in postganglionic nicotinic receptor sensitivity to ROS seen between sympathetic and ICG (parasympathetic) neurons.

By contrast, ganglionic transmission was resistant to the action of t-BHP. The distinct ROS generated by H2O2 and t-BHP is likely to underpin their different actions on ganglionic transmission, presumed to be at the presynaptic terminal. Indeed, t-BHP has been reported to increase field excitatory synaptic potentials underpinning the generation of long-term potentiation in the spinal cord (29).

The actions of H2O2 and t-BHP on the passive and active properties of the neurons of the intact neonatal and adult ICG preparation were broadly similar to those for dissociated neurons (55). Both ROS-generating agents caused a membrane hyperpolarization. This was larger for H2O2 than t-BHP, frequently resulting in the neuron becoming inexcitable. Similarly H2O2 hyperpolarized intestinal myenteric neurons in primary culture and hippocampal CA1 pyramidal neurons by inducing an increase in K+ conductance (37, 44, 51). In the present study, the rank order of the reversal of the H2O2 induced hyperpolarization by K+ channel blockers and muscarine (muscarinic ACh receptor agonist) was Ba2+ > TEA > muscarine > linopirdine, mirroring the inverse specificity of action of these agents.

There was no obvious difference between neurons displaying a transient depolarization and those in which it was absent (parameters examined included absolute resting Em, AHP50, Rin, and expression of TDR). Clearly, the presently available reasons underpinning this difference must remain speculative. Differences in membrane potential responses to ROS have also been reported for AH/type 2 myenteric neurons (52).

The H2O2-evoked hyperpolarization was fully reversed by TEA in neonates but only partially for adult ICG neurons. Differences in the expression of K+ channels in ICG neurons with postnatal development, for example the expression of the SK and M-channels (38, 39), may be responsible for the difference in action of this agent in neonates and adults.

ROS switched the evoked discharge characteristics of ICG neurons from predominantly phasic/multiple adapting to inexcitable/phasic. A similar change in firing pattern has been reported for the action of H2O2 on sympathetic ganglion neurons (18) and myenteric neurons (37). In dissociated sympathetic ganglion neurons, the alteration in evoked discharge was underpinned by oxidative modification and activation of M-type K+ channels. There is some indication that a similar mechanism may operate in ICG neurons.

The decrease in peak rates of rise and fall of the action potential are consistent with the action of ROS on Na+ and Ca2+ channels (55). These data suggest that H2O2-induced effects such as membrane potential hyperpolarization and neuronal silencing in ICG neurons are mediated by muscarine sensitive and delayed rectifier K+ channel activation. The lack of effect of XE991 and oxotremorine are likely due to their voltage-dependent action. This property underpins the limited actions of these agents in other whole mount preparations (40, 47).

The reduction of the hyperpolarization-activated nonspecific cation current (Ih) underlying TDR by H2O2 has previously been reported in myenteric neurons (51). Ih participates in the regulation of resting Em in ICG neurons (38, 39). A decrease in Ih will contribute toward the H2O2-evoked hyperpolarization of Em.

Catalase and SOD had distinct shielding effects against ROS donors on Em and ganglionic transmission. Catalase had a superior protective action compared with SOD against the action of H2O2, and this extended beyond the period of its application.

Catalase, in itself, had a small, but significant, hyperpolarizing action on ICG neurons. Perhaps it is abrogating the action of endogenous ROS. If so, endogenously produced ROS, at presumably low concentrations, has a depolarizing action. The distinct actions of low vs. high concentrations of H2O2 have been reported previously for synaptic transmission at the neuromuscular junction (20).

Both H2O2 and t-BHP increased resting intracellular Ca2+ levels. However, this increase was delayed until >15 min following application of H2O2. Catalase completely abrogated this H2O2-induced increase in [Ca2+]i. The simple explanation for these results is that endogenous ROS scavenger species are able to buffer the exogenously applied ROS to a limited extent. Provision of the exogenously applied ROS scavenger catalase will act to bolster the action of any available endogenous mechanisms.

The increase in resting [Ca2+]i could arise from several sources: Ca2+ release from ryanodine-sensitive stores, inhibition of Ca2+ uptake pumps in the endoplasmic reticulum and mitochondrial Ca2+ regulation inhibitors. Clearly, with presently available data, the source(s) must remain speculative. ROS inhibits voltage-gated Ca2+ channels in ICG neurons (55), so an increased influx through this route can be discounted. The rise in resting [Ca2+]i would seem to lag behind changes in the Em, confirming that the hyperpolarization is not secondary to a Ca2+-dependent mechanism. In AH/type 2 myenteric neurons the intracellular Ca2+ stores of the endoplasmic reticulum was proposed as the target for ROS (52). A recent report has demonstrated that the increased [Ca2+]i by H2O2 in rat myenteric neurons came from both intracellular and extracellular, transmembrane, sources. Furthermore, the extracellular flux of Ca2+ was blocked by Ca2+-dependent K+ channel blockers (37).

Perspectives and Significance

Impaired parasympathetic control of the heart is a powerful independent negatively prognostic predictor of arrhythmia and also a characteristic of myocardial infarction We have found that reactive oxygen species compromises the performance of ICG neurons. These ganglia represent the final common pathway through which the diverse, extrinsic neural signals to the heart are monitored before being transmitted to the effector tissues. Neuronal excitability of the postganglionic neuron was decreased by ROS. Ganglionic transmission was particularly sensitive to the actions of superoxide O2− and hydroxyl OH as indicated by the blocking actions of H2O2. Together, these actions will be detrimental to parasympathetic regulation of cardiac function and will produce a sympathovagal imbalance. Thus, there will be a predominance of sympathetic, proarrhythmic, activity. ROS scavengers mitigated the actions of ROS donors. Clearly, the development of strategies or interventions abrogating the blunting of ganglionic transmission in ischemia is important in the prevention of arrhythmia.

GRANTS

This work was supported by the British Heart Foundation (Project Grant PG/06/132/21753 to A. A. Harper).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.