H₂O₂ INDUCES DELAYED HYPEREXCITABILITY IN NUCLEUS TRACTUS SOLITARII NEURONS

Abstract

Hydrogen peroxide (H₂O₂) is a stable reactive oxygen species and potent neuromodulator of cellular and synaptic activity. Centrally, endogenous H₂O₂ is elevated during bouts of hypoxia-reoxygenation, a variety of disease states, and aging. The nucleus tractus solitarii (nTS) is the central termination site of visceral afferents for homeostatic reflexes and contributes to reflex alterations during these conditions. We determined the extent to which H₂O₂ modulates synaptic and membrane properties in nTS neurons in rat brainstem slices. Stimulation of the tractus solitarii (which contains the sensory afferent fibers) evoked synaptic currents that were not altered by 10 – 500 μM H₂O₂. However, 500 μM H₂O₂ modulated several intrinsic membrane properties of nTS neurons, including a decrease in input resistance, hyperpolarization of resting membrane potential (RMP) and action potential (AP) threshold (THR), and an initial reduction in AP discharge to depolarizing current. H₂O₂ increased conductance of barium-sensitive potassium currents, and block of these currents ablated H₂O₂-induced changes in RMP, input resistance and AP discharge. Following washout of H₂O₂ AP discharge was enhanced due to depolarization of RMP and a partially maintained hyperpolarization of THR. Hyperexcitability persisted with repeated H₂O₂ exposure. H₂O₂ effects on RMP and THR were ablated by intracellular administration of the antioxidant catalase, which was immunohistochemically identified in neurons throughout the nTS. Thus, H₂O₂ initially reduces excitability of nTS neurons that is followed by sustained hyperexcitability, which may play a profound role in cardiorespiratory reflexes.

The nucleus tractus solitarii (nTS) is vital to the cardiorespiratory reflex response during low arterial oxygen, blood pressure fluctuations, and other visceral stimuli (Andresen and Kunze, 1994; Kline et al., 2010). The nTS is the first central termination site in the brainstem for sensory afferents. nTS processing and integration of visceral sensory information is essential for an effective response to a given challenge, including chemoreflex augmentation of ventilation and sympathetic nerve activity in response to hypoxia (Andresen and Paton, 2011). Glutamate is the primary neurotransmitter released from afferent terminals, and numerous neuromodulators influence its release or postsynaptic receptor function (Kline et al., 2009; Sekizawa et al., 2009; Fawley et al., 2011), in addition to altering neuronal activity independent of neurotransmission. It is increasingly recognized that reactive oxygen species (ROS) are signaling neuromodulators in the cardiorespiratory system. ROS are implicated in the augmentation of sympathetic nerve activity in hypertension and heart failure (Ito et al., 2001; Zucker, 2006; Hirooka, 2008), in part due to actions within the nTS (Schultz et al., 2007; Wang et al., 2008).

Two major ROS are superoxide anion (O₂ ^·−) and the O₂^·− -derived, more stable hydrogen peroxide (H₂O₂; (Forman et al., 2010)). Due to its stability and ability to freely cross the plasma membrane, H₂O₂ has been suggested to be an important signaling molecule in synaptic and neuronal activity. H₂O₂ alters neurotransmission (Chen et al., 2001), synaptic long-term depression (Kamsler and Segal, 2003), ion channel function (Lebuffe et al., 2003; Avshalumov et al., 2005) and cellular activity (Avshalumov et al., 2005; Pouokam et al., 2009) in several central nuclei. The production and function of H₂O₂ is counter-balanced by cellular defense mechanisms (antioxidants), including catalase, which reduces H₂O₂ to H₂O and O₂ (Hou et al., 2010). H₂O₂, and its balance with antioxidants within the cardiorespiratory neuroaxis, modulates neuronal activity or function. For instance, in the rostral ventrolateral medulla of spontaneously hypertensive rats (SHR), hypertension is associated with elevated H₂O₂ and reduced catalase expression. Gene transfer of catalase into the rostral ventrolateral medulla reduces blood pressure in SHR (Chan et al., 2006). H₂O₂ in the hypothalamic paraventricular nucleus modulates regulation of sympathetic nerve activity (Cardoso et al., 2012). In the pre-Bötzinger complex, a region essential for respiratory rhythmogenesis, H₂O₂ initially depresses then augments action potential (AP) discharge and fictive breathing (Garcia et al., 2011). H₂O₂ in the nTS modulates blood pressure and heart rate control (Cardoso et al., 2009), and augments L-type calcium currents in isolated vagal-afferent labeled nTS cells (Wang et al., 2006). However, the function of H₂O₂ in modulating synaptic, neuronal or network activity in the nTS remains largely unknown.

We used electrophysiology in brainstem slices to determine the neurophysiological effect(s) of H₂O₂ on second-order nTS cells. Results demonstrate H₂O₂ hyperpolarized resting membrane potential (RMP) and initially reduced cellular excitability, but did not alter afferent-evoked synaptic transmission. The former effects are attributed to H₂O₂-mediated changes in membrane properties that are prevented by intracellular application of the antioxidant catalase. H₂O₂ also augmented conductance of barium-sensitive potassium currents, which were responsible for RMP hyperpolarization. Importantly, washout was associated with sustained hyperexcitability that persisted during subsequent application of H₂O₂. Last, immunohistochemistry identified catalase in neurons throughout the nTS. Altogether, these data demonstrate H₂O₂ is an important signaling molecule in the nTS.

Experimental Procedures

Ethical approval

The Animal Care and Use Committee of the University of Missouri approved all experimental protocols in accordance with NIH guidelines (“Guide for the Care and Use of Laboratory Animals”).

Animals

Male Sprague-Dawley rats (Harlan; n = 69, aged 4 – 6 weeks) were maintained in the AAALAC accredited vivarium of the Dalton Cardiovascular Research Center. The animals had water and food available ad libitum and were held at 22°C and 40% humidity on a 12 hour day/night cycle.

In Vitro Brainstem Preparation

As detailed previously (Kline et al., 2010), the brainstem of Isoflurane (VetOne) anesthetized rats was rapidly removed and placed in ice-cold low calcium-high magnesium aCSF (in mM: 124 NaCl, 3 KCl, 1.2 NaH₂PO₄, 1.2 MgSO₄, 25 NaHCO₃, 11 D-glucose, 0.4 L-Ascorbic Acid, 1 CaCl₂ and 2 MgCl₂, saturated with 95% O₂ - 5% CO₂, pH 7.4, ~300 mOsm). Horizontal slices (~280 μm) with lengthy segments of the tractus solitarii (TS) were cut using a vibrating microtome (VT 1000S, Leica). Tissue sections were placed in a superfusion chamber, secured via nylon mesh and superfused at ~3 mL/min with standard recording aCSF (in mM: 124 NaCl, 3 KCl, 1.2 NaH₂PO₄, 1.2 MgSO₄, 25 NaHCO₃, 11 D-glucose and 2 CaCl₂, saturated with 95% O₂ - 5%CO₂, pH 7.4, ~300 mOsm) at 31 – 33°C.

Electrophysiological recording

We utilized the patch clamp technique to record from nTS cell somas in the caudal portion of the medial nTS, a region that receives cardiorespiratory and gastrointestinal afferent information (Chitravanshi and Sapru, 1995; Guyenet, 2000; Travagli et al., 2006). Electrodes (King Precision Glass, type 8250) were pulled with a Flaming/Brown micropipette puller (Sutter Instruments, Model P-97) and had resistances of 3.0 – 4.5 MΩ when filled with standard recording solution (in mM: 10 NaCl, 130 K⁺ Gluconate, 11 EGTA, 1 CaCl₂, 10 HEPES, 1 MgCl₂, 2 MgATP, 0.2 NaGTP, pH 7.3, ~280 mOsm)(Kline et al., 2010). The recording pipette was guided with a piezoelectric micromanipulator (Burleigh, PCS-6000) and recordings were performed under voltage- and current-clamp. Neurons were rejected if the membrane potential was more positive than −45 mV upon initial rupture or if series resistance changed more than 20% throughout the experiment. In some experiments the series resistance was compensated. Liquid junction potential (approximately +13.2 mV) was not compensated. Data were recorded using a Multiclamp700B amplifier (Molecular Devices), filtered at 2 kHz and sampled at 20 kHz.

Protocols

Voltage clamp protocols

All neurons were clamped at −60 mV. Membrane conductance (i.e., input resistance, R_i) was determined from a 10 ms step-hyperpolarization to −65 mV. TS evoked EPSCs (TS-EPSCs) were generated with an isolated stimulator (World Precision Instruments, A310 Accupulser) and a concentric bipolar stimulating electrode (F. Haer) placed on the afferent containing TS. Stimulation intensity was increased until a TS-EPSC was evoked after which stimulation intensity was set at 1.5x TS-EPSC threshold. Properties of single TS-EPSCs at 0.5 Hz stimulation, or multiple (20) TS-EPSCs at 40 Hz stimulation [stimulus train; to mimic increased afferent discharge (Andresen and Kunze, 1994)] were measured in the absence and presence of H₂O₂ (10 – 500 μM). Initial studies using 1 mM H₂O₂ resulted in cellular and slice death, thus this concentration was not studied further. Miniature (m)EPSCs were recorded without external stimulation with 1 μM TTX (Tocris) and 10 μM bicuculline methobromide (Tocris) added to the aCSF. The lack of AP discharge to current ramps (see below) at the beginning of the mEPSC protocol ensured events were not due to action potential-driven network activity. While methyl derivatives of bicuculline have been shown to block Ca²⁺-activated K⁺ channels (Seutin et al., 1997) and alter AP-dependent neurotransmitter release, bicuculline methobromide used in the current experiments should not influence our mEPSC and current ramp protocols due to the presence of the sodium channel blocker TTX and our negative holding currents (−100 to −50 mV), which are outside the range over which these currents are observed in the nTS (Moak and Kunze, 1993). Because H₂O₂ has been shown to target IRK and K_2P channels in other tissue (Lebuffe et al., 2003; Avshalumov et al., 2005; Kim et al., 2007), we identified possible changes in K⁺ conductance through these channels by membrane voltage ramps (−100 mV to −50 mV, 1 sec.) and steps (−100 mV to −50 mV, 500 ms, 10 mV per step) (Bayliss et al., 1997; Austgen et al., 2012). Slope conductance was measured on ramps between −95 to −70 mV and with steps between −90 to −70 mV. Slope conductance during aCSF (pre-H₂O₂), barium chloride (BaCl₂, Sigma; 100 μM), H₂O₂ and H₂O₂ + BaCl₂ was examined. The subtracted H₂O₂- and Ba²⁺-sensitive currents were identified by subtracting the obtained current (e.g., H₂O₂) from the treatment immediately before (e.g., barium or aCSF). In a different set of experiments, we evoked K⁺ currents between −80 and +80 mV (20 mV per step, 400 ms) to measure the current amplitude of Atype potassium channels and other voltage-sensitive K⁺ channels in response to H₂O₂. Transient (A-type) K⁺ current under both conditions was examined by determining the difference between the peak and steady state current (Kline et al., 2010).

Current clamp protocols

AP discharge was evoked by either step-depolarization (0 to +90 pA, +10 pA steps, 100 ms duration) or ramp-depolarization (−20 pA to +50, +100 or +200 pA; 1 s ramp). Maximum ramp current used varied depending on the excitability of the particular cell, since some cells only responded with AP discharge to currents greater than 50 pA from RMP (I=0). AP threshold was determined from the AP differential and identified as the onset of the steepest slope. R_i was determined under current clamp conditions from a −20 pA current step while holding the cell at approximately −60 mV. The delay in AP discharge (i.e. delayed excitation, DE) was tested in a subset of cells by depolarizing current (+50, +100 or +200 pA, depending on the excitability of the cell; 800 ms; holding potential −60 to −65 mV) that was preceded by step hyperpolarizations (0 to −50 pA, −10 pA steps, 800 ms) to relieve A-type K⁺ channels from inactivation (Schild et al., 1993). RMP was measured under I=0 condition prior to current injection.

Physiologically relevant concentrations of H₂O₂ (10, 100, 300 and 500 μM H₂O₂ (Schroder and Eaton, 2008)) were freshly prepared from stock (Certified ACS 30%, Fisher Scientific) and added to the aCSF immediately before use. We typically recorded neurons 1–2 cell layers deep and perfused H₂O₂ for 5 minutes to allow penetration into the slice and to compensate for dead space in the perfusion tubing. To exclude potential pH changes of the bath solution by H₂O₂, we determined the pH of the aCSF-H₂O₂ solution up to 1 M H₂O₂. The highest concentration used in this study (500 μM H₂O₂) did not alter the pH of aCSF (pH change 0.002 ± 0.011; n = 5; p = 0.86, t-test). The first significant change of aCSF pH was measured at a H₂O₂ concentration 200-fold greater than used in the present study. For some experiments the antioxidant catalase (500 Units/mL; bovine liver, Sigma) was added to the recording solution within the recording pipette.

Immunohistochemistry

Immunohistochemistry was performed as in our previous manuscripts (Austgen et al., 2009; Kline et al., 2010). Briefly, deeply Isoflurane anesthetized rats were transcardially perfused using 0.1 M PBS followed by 4% paraformaldehyde (Sigma). The brainstem was removed and cut in 30 μm coronal brain sections with a vibratome (VT 1000S, Leica). The sections were rinsed in PBS, followed by heat-induced epitope retrieval in a decloaking chamber (Biocare Medical) at 80 °C (30 min) and then blocked by 10% normal donkey serum (Millipore) in 0.3 % Triton-PBS. Tissue sections were incubated with primary antibodies against the antioxidant catalase (rabbit, 1:300, ab1877, Abcam) and Neuronal nuclear antigen (NeuN; neuronal marker; mouse, 1:500, MAB377, Millipore) for 24 hours. Subsequently sections were rinsed and incubated in 0.3 % Triton-PBS including the following secondary antibodies: Cy2-conjugated donkey anti-rabbit IgG (1:200; 711-225-152, Jackson Immuno) and Cy5-conjugated donkey anti-mouse IgG (1:200; 715-175-151, Jackson Immuno). Sections were mounted on gelatin coated slides, air dried and coverslipped with ProLong Gold with 4′,6-diamidino-2-phenylindole (DAPI, nuclear marker; Invitrogen). Incubation of one section without primary antibody served as negative control (no fluorescent staining visible). Specificity of the antibody for catalase and NeuN has been confirmed previously [catalase, (Clarke et al., 2009); NeuN (Al-Khater et al., 2008)]. In this study, specificity for catalase antibody was also confirmed in nTS tissue via western blots (see below).

Immunoreactivity (-IR) was visualized with a conventional epifluorescent microscope (BX51, Olympus), a digital monochrome camera (ORCA-ER, Hamamatsu) and a spinning disc confocal unit (Olympus). Appropriate filter sets and excitation wavelength were used to visualize the different fluorophores. Z-stacks (0.5 μM) were taken (same focal planes) for each fluorophore used. Images were post-processed using ImageJ (Version 1.45r, NIH) by adjusting contrast and brightness for clarity.

Western blot

Western blots were performed as described previously (Kline et al., 2007). Frozen nTS tissue from 3 rats was pooled, homogenized in RIPA buffer (1% NP- 40, 0.5% DOC, 0.1% SDS, 0.15M NaCl, 50mM Tris/HCl and 2.5mM EDTA), and complemented with protease inhibitors (Complete, mini-EDTA-free tablets; Roche). Samples were incubated on ice for 2 hours and then centrifuged at 14,000 × G for 30 min at 4°C. Protein concentration of the supernatant was measured by the Micro BCA method (Pierce, Rockford, IL). Twenty micrograms of protein were separated on 4 – 20% Tris-Glycine gel (BioRad) and transferred to a PVDF membrane. Primary antibody anti-catalase (1:300; rabbit polyclonal, Abcam) was used to immunoblot.

Data Analysis

Electrophysiological data were analyzed with pClamp10 (Molecular Devices), MiniAnalysis (Synaptosoft) and Microsoft Excel software. Statistical analyses were performed using SigmaPlot 12.0 (Systat Software). Only second-order nTS cells, those that directly receive input from TS-afferents, were analyzed for this study. A direct connection was verified based on a low variability (jitter) of TS-EPSC onset (< 300 μsec.; SD of 30 TS-EPSC latencies from shock artifact, 20 kHz acquisition) (Doyle and Andresen, 2001; Kline et al., 2002; Hisadome et al., 2010; Accorsi-Mendonca et al., 2011). Each data point for a given TS-EPSC was an average of 30 events. Spontaneous postsynaptic current (sPSC and mEPSC) detection was set at 5x the root-mean square noise level and events were manually confirmed. Within an individual cell, Kolmogorov-Smirnov two sample tests were used to examine the distribution of spontaneous current amplitude and interevent intervals between pre-H₂O₂ and H₂O₂. Group sPSC and mEPSC data are reported as the sample average for amplitude and interevent frequency. To distinguish asynchronous (a)EPSCs from sEPSCs, the ratio of the number of EPSCs after and before the 40 Hz stimulus train was generated (after/before; ten 500 ms segments each). Generally, “baseline” refers to the aCSF condition prior to any H₂O₂ perfusion, unless noted otherwise (e.g. barium-baseline). The following statistical analyses were performed, and their p-values are reported in the text. The effect of H₂O₂ on EPSC amplitude, AP discharge threshold, AP delay, amplitude, rise slope, and decay slope was compared within treatments using paired t-tests. Synaptic and neuronal properties among multiple treatments were compared to their baseline response by one-way repeated measures (RM) ANOVA (with Student-Newman-Keuls post hoc test). Multiple treatments were compared by 2-way RM ANOVA (with Student-Newman-Keuls post hoc or Fishers LSD test). Statistical analyses were performed on the total sample size, unless otherwise noted. In addition the percentage of cells that responded to H₂O₂ by a change of at least 10% is provided in the text. H₂O₂ responses were also normalized and expressed relative to their individual baseline responses (defined as “1”). Results are considered significantly different at p values ≤ 0.05. Data are presented as means ± SEM.

RESULTS

In this study only cells monosynaptically connected to the tractus solitarii were evaluated. Across the cells tested, TS stimulation evoked invariant EPSCs with a mean amplitude of 131.74 ± 11.87 pA, latency of 4.63 ± 0.17 ms and jitter of 158.61 ± 5.82 μsec (n = 114). This is consistent with recording from nTS neurons that are monosynaptically connected to TS afferents (Kline et al., 2002; Hisadome et al., 2010; Accorsi-Mendonca et al., 2011).

H₂O₂ does not alter afferent-induced neurotransmission

As shown in the representative example in Figure 1A, stimulating the TS (0.5 Hz, average of 30 sweeps) elicited TS-EPSCs in nTS cells. The TS-EPSC amplitude was not altered during perfusion of the nTS with 500 μM H₂O₂. Across the cells tested, TS-EPSC amplitudes during H₂O₂ (10, 100, 300, 500 μM) were similar to vehicle or the baseline response (“1”; Figure 1B). For example, TS-EPSC amplitude was comparable during baseline (104.1 ± 17.5 pA) and 500 μM H₂O₂ (102.7 ± 16.8 pA; n = 8; p = 0.81). Also, TS-EPSC decay time (τ_90-10%) was not altered with any concentration of H₂O₂ (e.g. 500 μM: baseline, 3.31 ± 1.00 ms vs. H₂O₂, 2.39 ± 0.60 ms; n = 5; p = 0.24). Following exposure to H₂O₂, in a subset of cells 10 μM CNQX (non-NMDA antagonist) was applied. CNQX reduced the TS-EPSC amplitude by 86%, indicating that such TS-evoked currents are primarily mediated by glutamatergic non-NMDA receptors (baseline, 117.5 ± 15.4 vs. CNQX, 15.3 ± 4.8; n = 7; p < 0.001).

Figure 1. Effects of H₂O₂ on synaptic transmission.

A. Example of TS-EPSCs in the absence and presence of H₂O₂ (500 μM). Note H₂O₂ did not alter TS-EPSCs. Example shown is the average of 30 TS-EPSCs. B. H₂O₂ did not alter TS-EPSC amplitude compared to the baseline response or vehicle across multiple H₂O₂ concentrations. Data are normalized to each group’s baseline, defined as 1. C. Group data showing that H₂O₂ did not alter the current amplitude across the 20 events (use-dependent depression; also see inset) or the time course of depression when mimicking increased sensory afferent activity by augmenting TS stimulation frequency (40 Hz). Box in the inset highlights the depression from the first to the second pulse used for calculating the paired pulse ratio. All cells were held at -60 mV. Numbers of observations are depicted in parentheses.

Increasing TS stimulation frequency to 40 Hz (20 pulses, 10 sweeps) mimics a brief increase of afferent activity and provides insight into synaptic release properties. As typical of the first nTS synapse, high frequency stimulation progressively decreased the amplitude of TS-EPSCs following the first event (use-dependent depression; see example recording in Figure 1C inset). Consistent with the above results, neither H₂O₂ (10, 100, 300 and 500 μM) nor vehicle altered the amplitude of the TS-EPSCs compared to the baseline response, or the magnitude of change from the first event. Figure 1C contains the averaged results for 500 μM H₂O₂ (n = 8). The paired pulse ratio (PPR; ratio of TS-EPSC2/TS-EPSC1; box in the inset of Figure 1C), which differentiates alterations in presynaptic release or postsynaptic receptor properties, was also not altered by H₂O₂ (500 μM: baseline, 0.47 ± 0.06 vs. H₂O₂, 0.54 ± 0.07; n = 9; p = 0.35) or vehicle.

Asynchronous (a)EPSCs occur directly after TS stimulation and are another indicator of presynaptic alterations in transmitter release (Kline et al., 2007; Peters et al., 2010). To examine the role of H₂O₂ on aEPSCs, we examined the number of spontaneous EPSCs after the stimulus train and compared it to that occurring before the train. Subsequently, the EPSC after/before ratio was generated. A ratio greater than 1 indicates aEPSCs after the stimulus train. aEPSCs were observed following the stimulus train with all H₂O₂ concentrations tested, however, they were not altered compared to baseline (e.g. baseline ratio, 1.74 ± 0.32 vs. 500 μM H₂O₂, 2.19 ± 0.50; n = 9; p = 0.11). Taken together, these data suggest that H₂O₂ does not alter TS-afferent evoked EPSCs.

H₂O₂ decreases nTS network activity without altering miniature (m)EPSCs

Spontaneous postsynaptic currents (sPSCs) represent network activity within the nTS circuitry of the slice (Fortin and Champagnat, 1993). In response to 10, 100, 300, and 500 μM H₂O₂, the amplitude of sPSCs was not altered compared to baseline in any of the cells tested (e.g. for 500 μM: baseline, 21.63 ± 3.82 pA vs. H₂O₂, 24.01 ± 5.11 pA; n = 9; p = 0.39). H₂O₂ at 10 – 300 μM did not alter sPSC frequency. However, sPSC frequency decreased with 500 μM H₂O₂ (baseline, 26.81 ± 3.87 Hz vs. 500 μM H₂O₂, 20.67 ± 3.88 Hz; p < 0.01; a significant decrease to 0.74 ± 0.08 normalized to baseline; n = 9; 89 % of the 9 cells decreased ≥ 10 %). This decrease in sPSC frequency with H₂O₂ was independent of the cell’s baseline sPSC frequency (R² = 0.173; n = 9; p = 0.27, Pearson correlation). In the present study, sPSCs are likely glutamatergic excitatory currents due to their recording at a holding potential (−60 mV) near the calculated reversal potential of chloride (−59 mV). In addition, CNQX reduced or eliminated sPSCs in 7 cells by 93.2 ± 5.6 % (p < 0.001), in agreement with our previous studies (Kline et al., 2010).

To further identify potential effects of H₂O₂ on synaptic transmission, we examined mEPSCs in the presence of TTX (1 μM; a voltage-gated Na⁺-channel blocker) and bicuculline (10 μM; a GABA_A-receptor antagonist). Given that lower concentrations of H₂O₂ did not alter synaptic parameters, we examined 500 μM H₂O₂ exclusively. Compared to baseline, 500 μM H₂O₂ did not change mEPSC amplitude (baseline, 35.58 ± 5.4 pA vs. 500 μM H₂O₂, 36.32 ± 5.2 pA; n = 7; p = 0.77) or frequency (baseline, 10.56 ± 4.08 Hz vs. 500 μM H₂O₂, 12.69 ± 5.28 Hz; n = 7; p = 0.23). Thus, H₂O₂ does not alter mEPSC amplitude or frequency in the absence of network activity.

H₂O₂ hyperpolarizes resting membrane potential (RMP) and action potential threshold (THR)

We tested the extent to which H₂O₂ alters cell membrane properties. Changes in input resistance (R_i) are indicative of an alteration in channel conductance. Perfusions with vehicle and 10, 100 and 300 μM H₂O₂ did not alter R_i. By contrast, 500 μM H₂O₂ significantly and reversibly decreased R_i, from 485.35 ± 46.00 MΩ (baseline) to 367.42 ± 31.36 MΩ (p < 0.05, n = 9; 78 % of the 9 cells decreased by 10 % or more). During washout, R_i was 509.09 ± 95.91 MΩ (n = 5). nTS cells had a mean RMP of -56.75 ± 0.87 mV (n = 122) across all cells tested. Relative to baseline and vehicle, H₂O₂ at 10 and 100 μM did not alter RMP, yet 300 and 500 μM H₂O₂ significantly hyperpolarized RMP (Figure 2A). The raw values for RMP and their changes with H₂O₂ are shown in Table 1. This effect typically occurred between 1 – 3 minutes after the start of H₂O₂ perfusion. THR of all nTS cells tested was −28.34 ± 0.58 mV (n = 93). Relative to baseline and vehicle, 10 – 300 μM H₂O₂ did not alter THR but 500 μM H₂O₂ significantly hyperpolarized THR (Figure 2B). The raw values for THR and their changes with H₂O₂ are shown in Table 1. Since only 500 μM H₂O₂ significantly altered R_i, RMP and THR, we chose this concentration for the remainder of this study to examine the effects of H₂O₂.

Figure 2. H₂O₂ alters neuronal membrane properties.

RMP (A) and THR (B) relative to their baseline (1 = no change, >1 = hyperpolarization) for vehicle, 10 – 500 μM H₂O₂ and 500 μM H₂O₂ + 500 U/mL catalase in the pipette. H₂O₂ did not alter RMP or THR at 10 and 100 μM. However, 300 and 500 μM H₂O₂ significantly hyperpolarized RMP, and 500 μM H₂O₂ significantly hyperpolarized THR. Further, the changes mediated by 500 μM H₂O₂ in RMP and THR were eliminated by catalase in the recording pipette. Two-way RM ANOVA; * compared to baseline, ‡ compared to vehicle, 10 μM and 500 μM H₂O₂ + 500 U/mL catalase for THR, and additionally to 100 μM H₂O₂ for RMP; p < 0.05. Numbers of observations are indicated in parentheses. CAT = catalase.

Table 1. RMP and THR is altered by H₂O₂ and blocked by exogenous catalase.

Raw values of the dose-response relationship between different H₂O₂ concentrations and RMP and THR. Note, only 500 μM H₂O₂ significantly altered both parameters, which was antagonized by increasing the intracellular antioxidant catalase.

Resting membrane potential (RMP) [mV]	baseline	treatment	n	p
vehicle	−57.00 ±4.79	−56.58 ±4.56	7	0.79
10 μM H2O2	−55.70 ±3.34	−53.77 ±3.93	11	0.32
100 μM H2O2	−53.50 ±2.16	−52.77 ±1.87	12	0.63
300 μM H2O2	−61.07 ±2.76	−65.39 ±2.43	9	0.07
500 μM H2O2	−64.83 ±2.36	−71.10 ±2.23	9	0.04
500 μM H2O2 + catalase	−66.88 ±2.48	−66.55 ±2.57	8	0.79

Action potential threshold (THR) [mV]	baseline	treatment	n	p
vehicle	−27.69 ±1.14	−28.90 ±1.73	9	0.18
10 μM H2O2	−30.82 ±1.27	−30.78 ±2.20	12	0.98
100 μM H2O2	−34.45 ±2.40	37.36 ±3.84	5	0.16
300 μM H2O2	−33.23 ±2.50	−36.63 ±2.01	5	0.04
500 μM H2O2	−28.70 ±1.39	−34.70 ±3.04	9	0.02
500 μM H2O2 + catalase	−28.69 ±1.95	−29.57 ±2.10	7	0.10

Statistical analysis using paired t-tests; p value shown on the right. “n” represents the number of observations.

When effects of 500 μM H₂O₂ on RMP and THR were examined in a larger dataset (Table 2), there also was a significant hyperpolarization of RMP (62 % of cells hyperpolarized ≥ 10 %) and THR (74 % of cells hyperpolarized ≥ 10 %). The magnitude of response for RMP and THR to 500 μM H₂O₂ was similar between the cells examined in the initial dose-response (n = 9) and this population (p = 0.35 [RMP] and p = 0.48 [THR]). With washout, RMP returned to baseline whereas THR remained significantly hyperpolarized. The difference between THR and RMP ([THR-RMP]), which is the voltage that must be exceeded by a depolarizing stimulus to initiate AP discharge, was not affected during H₂O₂ perfusion, but decreased significantly at washout. Rheobase significantly increased from baseline to H₂O₂, and decreased significantly during wash. Other membrane properties of the first induced AP during baseline, H₂O₂ and wash are contained in Table 2.

Table 2. Membrane potential properties during baseline and H₂O₂.

Properties of THR, RMP, [THR-RMP], rheobase and other parameters of the first AP generated by step depolarization during baseline, H₂O₂ and wash.

	baseline	H2O2	wash	n (H2O2/wash)
RMP [mV]	−59.82 ±1.56	−69.61 ±2.36 *	−57.17 ±1.83 †	26/23
THR [mV]	−26.09 ±1.03	−33.35 ±2.15 *	−31.13 ±1.58 *	23/20
[THR-RMP] [mV]	33.33 ±1.95	35.62 ±1.83	25.74 ±1.86 *†	23/20
Rheobase [pA]	66.55 ±13.55	92.15 ±15.58 *	38.40 ±6.69 †	23/19
Overshoot [mV] (over 0 mV)	35.61 ±1.71	25.10 ±3.15 *	26.60 ±3.21 *	23/20
AP amplitude from RMP [mV]	95.03 ±2.20	94.07 ±2.72	83.47 ±3.61 *†	23/20
AP amplitude from THR [mV]	61.70 ±1.64	58.45 ±2.25 *	57.73 ±3.00	23/20
Half width [ms]	1.11 ±0.06	1.17 ±0.05	1.19 ±0.05	19/16
Rising slope [mV/ms]	195.39 ±11.47	178.70 ±13.31	158.08 ±13.80 *	20/17
Falling slope [mV/ms]	−93.16 ±6.05	−86.64 ±4.53	−72.49 ±4.21 *†	20/17
Afterhyperpolarisation from THR [mV]	19.94 ±3.74	18.72 ±4.79	15.51 ±4.14	23/20

One-way RM ANOVA;

^*compared to baseline,

^†compared to H₂O₂; p < 0.05. “n” represents the number of observations during H₂O₂ and wash.

H₂O₂ may alter RMP and THR by modifying proteins through extracellular or intracellular domains. To determine if H₂O₂ was acting intracellularly, membrane-impermeable catalase was included within the pipette solution and effects of H₂O₂ were reevaluated. Catalase is a primary antioxidant of the cell that specifically breaks down H₂O₂ into H₂O and O₂. Adding 500 Units/mL catalase into the recording pipette antagonized the hyperpolarizing effects of 500 μM H₂O₂ on RMP and THR. The lack of change in RMP and THR during 500 μM H₂O₂ + 500 Units/mL catalase from its baseline as well as compared to other H₂O₂ concentrations is shown in Figure 2A, B and Table 1. These data indicate an intracellular site of modulation by H₂O₂ and suggest that the changes in RMP and THR are mediated specifically by H₂O₂.

H₂O₂ modulates the firing rate (AP discharge) of nTS neurons

Figure 3A shows a representative example of the effects of 500 μM H₂O₂ on the neuronal properties of an nTS cell. Step current depolarization (0 to +40 pA, +10 pA steps, 100 ms duration) evoked AP discharge under baseline conditions. As described above, H₂O₂ induced membrane hyperpolarization and upon current depolarization, there was a decrease in AP discharge and overshoot. In addition, the THR voltage for evoking an AP was shifted in the hyperpolarized direction. With five minutes of washout, RMP returned to baseline level and upon current depolarization, AP discharge again occurred. Furthermore, compared to the baseline response, discharge was greater following washout.

Figure 3. H₂O₂ induced changes in neuronal excitability.

A. Representative recording of the membrane potential during step depolarization. Left, Progressive depolarization (inset = stimulus waveform; 10 pA steps, 100 ms) induced AP discharge from RMP. Middle, Bath application of H₂O₂ (5 min; 500 μM) hyperpolarized RMP. In addition, THR was hyperpolarized and AP discharge was reduced. Right, Wash (5 min following H₂O₂) reversed the H₂O₂-induced change in RMP but not THR. The cell became hyperexcitable during depolarization. B. H₂O₂-induced reduction in R_i. Note the partial reversal following wash. C. Mean data for RMP, THR and their difference [THR-RMP] for baseline, 500 μM H₂O₂ and wash. Note the hyperpolarization of THR and RMP during H₂O₂, and the reduced difference between THR and RMP ([THR-RMP]) during wash, which is the voltage to be overcome by a depolarizing stimulus. D. Plot of AP discharge to increasing current step depolarization. Data are normalized to the maximum AP discharge under the baseline (bsl) condition which was defined as 1. H₂O₂ decreases AP discharge to step depolarization. This effect is readily reversible at wash and AP discharge becomes significantly greater (i.e., hyperexcitable). B & C, one-way RM ANOVA; D, two-way RM ANOVA; * compared to the baseline response, † compared to H₂O₂; p < 0.05. Numbers of observations are depicted in parentheses.

We quantified the input-output relationship of AP discharge in a subset of 13 cells, for which also the responses of R_i, RMP and THR to 500 μM H₂O₂ are shown in Figure 3. As in the previous experiments, H₂O₂ decreased R_i (Figure 3B; 92 % of these cells decreased by ≥10 %). With wash, R_i partially returned to baseline and was significantly different from H₂O₂. H₂O₂ significantly hyperpolarized RMP (Figure 3C bottom) and THR (Figure 3C middle) as observed in the large dataset above. Upon wash, in this subpopulation RMP returned again to baseline, but THR remained hyperpolarized, which significantly reduced [THR-RMP] (Figure 3C top). Quantifying the AP discharge in response to step depolarizations (0 to +90 pA, +10 pA steps, 100 ms duration, corresponding to our observed TS-EPSC amplitudes), there was a progressively greater discharge during baseline (Figure 3D, with the maximal discharge for each individual cell plotted as 1). The prominent decrease in R_i during H₂O₂ significantly increased rheobase from 31.5 ± 4.5 pA (baseline) to 67.7 ± 16.9 pA (H₂O₂, n = 13, p ≤ 0.05). This significantly reduced AP discharge evoked by depolarizing currents ≥ 50 pA when compared to the baseline response. At wash, rheobase significantly decreased to 32.3 ± 6.8 pA (n = 13, p ≤ 0.05). This returned AP discharge to baseline values that even significantly exceeded baseline discharge upon depolarization ≥ 60 pA. Taken together, these results suggest H₂O₂ initially decreases AP discharge, but induces hyperexcitability upon washout when compared to the baseline response.

To further examine and correlate H₂O₂-induced alterations in R_i, RMP and THR to neuronal discharge, we monitored these properties in continuous current clamp during repetitive H₂O₂ applications. Membrane potential during current ramps in response to repetitive exposure to vehicle (aCSF) or H₂O₂ (two treatments with 500 μM H₂O₂, 10 min. each) are shown in Figure 4A & B. Vehicle did not alter AP discharge over the 40 minute recording time (T_5min – T_40min). In contrast to vehicle and similar to above, a single bout of H₂O₂ decreased AP discharge relative to its baseline response. Washout increased discharge above baseline values that persisted even following a subsequent H₂O₂ perfusion and washout. Quantitatively, AP discharge to current ramps was significantly higher subsequent to the first H₂O₂ exposure when compared to vehicle or the baseline response (Figure 4C). THR, RMP, [THR-RMP], R_i and rheobase during repetitive H₂O₂ and vehicle are quantified in Table 3. While THR tended to hyperpolarize over time during vehicle (T_5min – T_40min), it was not significantly different from baseline. Likewise R_i, RMP, [THR-RMP] and rheobase were unaltered in vehicle. On the contrary to vehicle and similar to the protocol with one bout, in this sample H₂O₂ initially (T_10min) significantly hyperpolarized RMP, tended to decrease R_i, did not change THR (note, only 4 cells that continued to respond), and tended to increased rheobase. During and following the first washout, there is a significant reduction in [THR-RMP] and rheobase, correlating with the increased AP discharge shown in Figure 4C. After 40 minutes of recording (wash/T_40min) data for H₂O₂ treatment shows a significantly depolarized RMP and hyperpolarized THR, explaining the observed decreases in [THR-RMP] and rheobase, and likely resulting in the increase in AP discharge. Thus, although the protocol was designed to evaluate the effects of repeated H₂O₂ application, the excitatory effects of the initial washout persisted throughout the second application of H₂O₂.

Figure 4. Prolonged hyperexcitability due to H₂O₂.

Representative traces of AP discharge during ramp depolarization at repetitive vehicle (A) or H₂O₂ perfusion and wash (B). The time points shown (T_5min – T_40min) during vehicle and H₂O₂ depict the time at which a ramp depolarization was performed, which was 5 minutes after a new treatment. Note the AP discharge to depolarization during baseline, which was eliminated with H₂O₂, but was augmented following washout and persisted during a second bout of H₂O₂. Vehicle did not alter AP discharge over the 40 minute period. C. Group data for AP discharge during ramp depolarization. Data are presented as the relative increase from each cell’s baseline response (normalized to “1”). Note the significant hyperexcitability following the initial H₂O₂ perfusion, when compared to the baseline response or vehicle (time control). *#, p < 0.05, two-way RM ANOVA; * compared to baseline, # compared to vehicle. Numbers of observations are depicted in parentheses.

Table 3. Membrane properties during repetitive exposure to H₂O₂.

Membrane properties (raw and normalized to their baseline [defined as “1”]) during baseline, two bouts of H₂O₂ or vehicle and their respective wash. Top panel, vehicle; bottom panel, H₂O₂ (500 μM). T_5min – T_40min represent the time points during H₂O₂ or vehicle in which cell properties were measured. Note THR, RMP, [THR-RMP], R_i and rheobase did not change during vehicle when compared to the respective baseline values. By contrast, in this smaller cell sample, the initial bout of H₂O₂ hyperpolarized RMP and reduced R_i in the same direction as our larger sample. H₂O₂ significantly altered RMP, THR, [THR-RMP], and rheobase when compared to baseline values at later time points.

		baseline/T5min	treatment/T10min	wash/T20min	treatment/T30min	wash/T40min	n
VEHICLE
THR	raw [mV]	−28.28 ±2.65	−29.75 ±3.01	−32.30 ±2.89	−32.72 ±3.22	−32.51 ±3.49	6
	normalized	1.00 ±0.00	1.05 ±0.02	1.15 ±0.03	1.16 ±0.03	1.14 ±0.04
RMP	raw [mV]	−50.99 ±4.03	−50.46 ±3.25	−50.92 ±2.97	−51.38 ±2.94	−51.55 ±2.95	6
	normalized	1.00 ±0.00	1.00 ±0.02	1.01 ±0.04	1.02 ±0.06	1.02 ±0.05
[THR-RMP]	raw [mV]	22.72 ±4.09	20.71 ±3.19	18.62 ±3.25	18.66 ±3.22	19.04 ±3.08	6
	normalized	1.00 ±0.00	0.95 ±0.06	0.85 ±0.08	0.88 ±0.12	0.91 ±0.13
Ri	raw [mΩ]	628.01 ±104.29	632.15 ±110.96	649.18 ±134.14	655.46 ±126.27	707.62 ±132.96	6
	normalized	1.00 ±0.00	1.02 ±0.05	1.02 ±0.10	1.04 ±0.11	1.24 ±0.24
Rheobase	raw [pA]	21.08 ±6.49	18.01 ±4.49	14.04 ±4.64	20.65 ±9.50	14.98 ±3.54	6
	normalized	1.00 ±0.00	0.93 ±0.17	0.85 ±0.22	1.12 ±0.25	1.02 ±0.30
H2O2
THR	raw [mV]	−26.48 ±1.07	(−24.62 ±1.83)	−28.57 ±2.44	−31.63 ±4.50 *	−30.65 ±3.20 *	(4),7
	normalized	1.00 ±0.00	(0.95 ±0.03)	1.07 ±0.05	1.17 ±0.12 *	1.14 ±0.08 *
RMP	raw [mV]	−63.34 ±0.89	−68.34 ±4.54	−59.46 ±2.06	−60.30 ±4.35	−54.36 ±2.80 *	7
	normalized	1.00 ±0.00	1.08 ±0.06	0.94 ±0.03	0.95 ±0.06	0.86 ±0.04 *
[THR-RMP]	raw [mV]	36.86 ±0.97	(38.05 ±3.10)	30.89 ±2.64 *	28.66 ±2.51 *	23.70 ±2.70 *	(4),7
	normalized	1.00 ±0.00	(1.03 ±0.06)	0.84 ±0.07 *	0.77 ±0.06 *	0.64 ±0.07 *
Ri	raw [mΩ]	609.47 ±91.9	467.36 ±105.44	587.34 ±88.17	550.19 ±96.29	561.21 ±128.60	7
	normalized	1.00 ±0.00	0.76 ±0.13	0.98 ±0.06	0.89 ±0.07	0.88 ±0.10
Rheobase	raw [pA]	73.52 ±13.73	(78.87 ±29.79)	55.53 ±15.57 *	53.49 ±14.69 *	40.66 ±8.63 *	(4),7
	normalized	1.00 ±0.00	(1.32 ±0.20)	0.75 ±0.12	0.68 ±0.10	0.57 ±0.09

Two-way RM ANOVA; comparisons shown within groups;

^*p < 0.05; compared to baseline. “n” represents the number of observations at each time point. Note three cells did not discharge during the initial bout of H₂O₂ so that certain parameters could not be measured; values in parentheses indicate the number of cells that continued to discharge during the 1^st H₂O₂.

H₂O₂ modulates potassium currents

H₂O₂ initially hyperpolarizes RMP and reduces R_i suggesting the involvement of one or more K⁺-currents that operate at RMP. We postulated the involvement of IRK and K_2P currents which have been shown in other systems to control membrane potential and be altered by H₂O₂. Here we compared holding currents during voltage ramps (−100 mV to −50 mV, 1 sec.) within the conducting range of IRK and K_2P channels (Bayliss et al., 1997; Day et al., 2005), under baseline and H₂O₂ perfusion. Prior to application of the voltage ramp, H₂O₂ produced a general positive (outward) shift in holding current consistent with hyperpolarization of RMP (example from one cell shown in Figure 5A). The voltage ramp during baseline elicited an inward holding current with an average slope conductance of 2.222 ± 0.522 pA/mV (nS, measured between −95 to −70 mV; n = 6). The addition of H₂O₂ significantly increased slope conductance to 2.716 ± 0.625 nS (n = 6; p < 0.05; an increase to 1.27 ± 0.12 when normalized to the baseline response), indicating an increase of IRK/K_2P channel conductance. The subtracted H₂O₂-sensitive current reversed at −75.4 ± 5.1 mV (n = 6; Figure 5B), towards our calculated reversal potential of potassium (−98.7 mV). In five cells, adding 100 μM barium chloride (BaCl₂), a concentration which blocks IRK and K_2P channels (Day et al., 2005), to the H₂O₂ perfusion solution significantly reduced slope conductance from 2.716 ± 0.625 nS (during H₂O₂) to 0.800 ± 0.073 nS (p < 0.05). This Ba²⁺-sensitive current reversed at −81.2 ± 1.5 mV, which again is towards the calculated equilibrium potential for K⁺ in our recording conditions and not significantly different than the H₂O₂ sensitive current. These data show that H₂O₂ increases K⁺ conductance, which is subsequently antagonized by barium. The H₂O₂-effect on increased slope conductance during voltage ramps was additionally confirmed by a conventional voltage step protocol between −100 and −50 mV. Similarly, addition of H₂O₂ significantly increased the IRK/K_2P channel conductance (slope measured between −90 to −70 mV steps; baseline, 1.785 ± 0.347 nS vs. H₂O₂, 3.098 ± 0.412 nS; n = 6; p ≤ 0.01; an increase to 1.95 ± 0.34 when normalized to the baseline response). To eliminate potential effects of tonic background currents from ionotropic receptors on our responses, we repeated H₂O₂ perfusion during the voltage step protocol in the presence of the non-NMDA receptor antagonist CNQX (10 μM) and bicuculline methobromide (10 μM). During H₂O₂, cells demonstrated the typical increase in holding current (i.e. hyperpolarization of RMP) and decrease of R_i, and increased conductance through IRK/K_2P channels (slope measured between −90 to −70 mV steps; CNQX + bicuculline-baseline, 1.728 ± 0.271 nS vs. H₂O₂, 2.307 ± 0.159 nS; n = 5; p ≤ 0.05). An example is shown in Figure 5C & D.

Figure 5. H₂O₂ increases the conductance through potassium channels.

Examples of the holding currents recorded during voltage ramps (A–B, E–F) and voltage steps (C–D). A. H₂O₂ (dark grey trace) elicited a positive shift in holding current at −60 mV when compared to baseline (black trace), as shown in the segments at the beginning and end of each trace (dotted line denotes initial baseline holding potential). H₂O₂ also increased the slope conductance of holding currents during voltage ramps. Adding 100 μM BaCl₂ to H₂O₂ (light grey trace) reversed the H₂O₂-induced increase in slope conductance. Note the breakthrough sodium spikes towards the end of the ramp, a region not included in our slope conductance measurements. Inset, Voltage ramp used to examine the holding currents from −100 to −50 mV (1 sec. ramp). B. The subtracted H₂O₂-sensitive current. C. Verification of the observed H₂O₂-effects using voltage steps from −100 to −50 mV (10 mV steps) during perfusion of CNQX and bicuculline (Bic) to block potential tonic background and synaptic currents. H₂O₂ again elicited a positive shift in holding current (i.e. RMP hyperpolarization) illustrated by the initial segments before the step protocol. During H₂O₂ the conductance increased with each voltage step. D. Current-voltage relationship. H₂O₂ increased slope conductance, similar to panel A, in response to voltage steps. E. Barium (light grey trace) elicited a negative shift in holding current at -60 mV when compared to baseline (black trace). Barium decreased the slope conductance of holding currents to the voltage ramps. Adding H₂O₂ to barium chloride (dark grey trace) induced no additional changes. F. The subtracted barium-sensitive and H₂O₂-sensitive current. Note that the H₂O₂-sensitive current is completely eliminated in the presence of barium.

Next we determined whether pre-block of Ba²⁺-sensitive currents would prevent the increase in conductance by H₂O₂. Barium alone produced a negative shift in holding current under voltage clamp condition (baseline, −18.23 ± 8.76 pA vs. BaCl₂, −23.28 ± 9.40 pA; n = 6; p < 0.05, Figure 5E), consistent with depolarization of RMP (see below). In the presence of BaCl₂, H₂O₂ perfusion had no effect on holding current (−23.01± 9.19 pA; n = 6; p = 0.89). As expected, slope conductance during voltage ramps significantly decreased from baseline to barium (baseline, 1.827 ± 0.478 nS vs. BaCl₂, 1.316 ± 0.374 nS; n = 6; p < 0.05), indicating blockade of IRK/K_2P channel conductance. The subtracted Ba²⁺-sensitive current reversed at −79.7 ± 2.3 mV (example in Figure 5F), similar to the above results. Barium prevented the H₂O₂-induced increase in slope conductance (1.242 ± 0.334 nS; n = 6; p = 0.30). The elimination of the subtracted H₂O₂-sensitive current is shown in Figure 5F. These results have also been affirmed in voltage step protocols where barium prevented increases in conductance due to H₂O₂ (BaCl₂, 1.084 ± 0.170 nS vs. H₂O₂ and BaCl₂, 1.014 ± 0.185 nS; n = 6; p = 0.48). Altogether, these results confirm that H₂O₂ alters potassium conductance, likely through of IRK/K_2P channels.

H₂O₂-sensitive potassium currents modulate RMP, R_i and AP discharge

To examine whether H₂O₂-sensitive IRK/K_2P channels are responsible for membrane hyperpolarization, we monitored RMP during H₂O₂ and BaCl₂ (Figure 6). RMP significantly depolarized with Ba²⁺, and H₂O₂ produced no additional change in the presence of Ba²⁺ (Figure 6A). Thus, barium (presumably by blocking IRK and K_2P channels) eliminated the effect of H₂O₂ on RMP. THR, on the other hand, remained stable with Ba²⁺ perfusion but significantly hyperpolarized with the addition of H₂O₂. The barium-dependent depolarization of RMP significantly decreased [RMP-THR], which however remained stable during the H₂O₂-dependent hyperpolarization of THR. Barium also blocked the H₂O₂-dependent reduction in R_i and AP discharge, as well as the increase in rheobase (Figure 6B–D). Overall, these data suggest that changes in RMP and R_i, and the resulting decrease in AP discharge by H₂O₂ are due to increases in IRK and K_2P conductance.

Figure 6. Potassium channel blockade with barium eliminates H₂O₂-induced changes in membrane properties.

Membrane properties and AP discharge were measured before and during 1 sec ramp depolarizations. Note, barium blocked H₂O₂-dependent alterations in RMP (A) and R_i (B), and the resulting changes in [THR-RMP], rheobase (C) and AP discharge (D). Barium did not prevent H₂O₂-induced hyperpolarization of THR. One-way RM ANOVA, * compared to baseline response, † compared to barium; p < 0.05. Numbers of observations are depicted in parentheses.

H₂O₂ and delayed excitation (DE) of nTS neurons

nTS neurons exhibit heterogeneity in their firing properties, namely “phasic”, “tonic” and “delayed excitation” cells (Paton et al., 1993; Kline et al., 2010). We examined the extent to which H₂O₂ may alter one or more of these firing phenotypes in a subset of cells. Cellular phenotype was classified by discharge properties in response to a depolarizing current that was preceded by hyperpolarization (0 to −50 pA, −10 pA steps, 800 ms duration; inset Figure 7A). Out of 12 nTS neurons examined for their firing properties, 8 neurons were DE cells, 3 neurons were tonic firing cells and 1 cell exhibited phasic firing. In the small subset of tonically or phasic firing cells, discharge delay was not altered by either hyperpolarizing pre-pulses or H₂O₂, and thus these cells were not studied further.

Figure 7. H₂O₂ alters delayed excitation of nTS neurons.

A. Representative traces of AP discharge of cells characterized by delayed excitation (DE). Left. During baseline, a hyperpolarizing pre-pulse (grey trace) increased AP discharge delay to a depolarizing stimulus, compared to no pre-pulse (black trace). Arrows depict the initial AP for each condition. See inset for stimulus. Right. During H₂O₂, while maintaining the cell at baseline (bsl) membrane potential using bias current (i.e., preventing H₂O₂-induced hyperpolarization), an increase in AP discharge to depolarization is observed (without pre-pulse; black trace). This is associated with the reduction in THR. A hyperpolarizing pre-pulse significantly delayed AP discharge (grey arrow), similar to the baseline response. B. Group data plotting pre-pulse potential (x-axis; elicited by hyperpolarizing current) vs. AP discharge delay (y-axis) under baseline (black line) and H₂O₂ (grey line) for cells exhibiting DE (n = 5 – 8, delayed spiking of some cells was greater than the duration of the depolarizing pulse when pre-hyperpolarization was ≥ −20 pA). Note the overall reduced delay in AP discharge with H₂O₂ during step hyperpolarization. Also note that increasing the hyperpolarizing pre-pulse increased the delay during both the baseline response and H₂O₂, and this increase in delay was not altered by H₂O₂. For comparison, regression lines correspond to only pre-pulse potential values more depolarized than −90 mV. *, averaged delay of multiple pre-pulse potentials, baseline vs. H₂O₂, p < 0.05, t-test.

A representative example and the role of H₂O₂ on a DE cell is shown in Figure 7A. As typical of a DE nTS cell, during aCSF baseline, onset of AP discharge to a depolarizing pulse was delayed (black arrow denotes initial spike). Augmenting hyperpolarization increased the delay in discharge upon depolarization (grey arrow denotes initial spike, Figure 7A). Such delay reflects hyperpolarization-induced relief of inactivation of A-type K⁺-currents that enables their activation upon depolarization (Schild et al., 1993). Preventing the hyperpolarization of RMP by H₂O₂ via bias current (−65 mV; Figure 7A right) decreased the initial delay in AP discharge in response to the depolarizing pulse (black arrow). Hyperpolarization in the presence of H₂O₂ induced AP delay (DE, grey arrow) when compared to no prior hyperpolarization (black arrow).

The group data illustrating the effect of H₂O₂ on DE in nTS neurons are shown in Figure 7B. As shown in our representative example (Figure 7A), there was an initial delay in depolarization-induced AP discharge, and increasing hyperpolarization prior to depolarization significantly prolonged the delay of AP discharge under baseline and H₂O₂ (Figure 7B; p < 0.001). Due to the H₂O₂-induced decrease in R_i, negative current injection did not hyperpolarize the membrane to the same potential as during the baseline response. Thus, we compared AP delay between baseline and H₂O₂ only under similar hyperpolarizing pre-pulse potentials (i.e. depolarized more than −90 mV). During H₂O₂, the overall delay in discharge was reduced compared to baseline (average delay for all pre-pulse potentials > −90 mV; baseline, 189.15 ± 41.83 ms vs. H₂O₂, 123.13 ± 36.82 ms; n = 8; p < 0.05). However, there was a similar progression of the delay (based on the regression) during baseline and H₂O₂ (average slope of regression > −90 mV; baseline, −6.82 ± 1.52 vs. H₂O₂, −6.04 ± 2.11; n = 8; p = 0.71). Such results may suggest A-type K⁺-current amplitude is unaltered by H₂O₂. This was confirmed by examining voltage-activated outward K⁺-currents in response to voltage steps (−80 to +80 mV; 400 ms; 20 mV steps). There was a small but non-significant decrease in transient K⁺ currents by H₂O₂ at voltages greater than 0 mV (n = 9; p = 0.31).

An important part of this protocol held the initial RMP constant (−60 to −65 mV without a hyperpolarizing pre-pulse) during baseline and H₂O₂ with bias current. This was done to examine the influence of H₂O₂ on AP discharge in the absence of membrane hyperpolarization. Under these conditions, H₂O₂ increased overall AP discharge in response to current depolarization. Elevated discharge was observed without pre-hyperpolarization steps (i.e., black trace in Figure 7A right) in 5 out of 8 cells (baseline, 9.2 ± 3.1 vs. 500 μM H₂O₂, 22.8 ± 4.1 AP/stimulus; n = 5; p < 0.01). In the remaining cells, AP discharge either decreased (1 cell) or was stable (2 cells). The data across this entire population did not reach statistical significance (baseline, 8.25 ± 2.01 vs. 500 μM H₂O₂, 15.38 ± 4.40 AP/stimulus; n = 8; p = 0.09). In the presence of bias current, H₂O₂ significantly hyperpolarized THR across all DE neurons (THR at 0 pA hyperpolarization: baseline, −26.49 ± 1.96 mV vs. 500 μM H₂O₂, −35.04 ± 3.67 mV; n = 8; p < 0.01).

Catalase in nTS neurons

H₂O₂ modulates cellular excitability which was antagonized by intracellular application of catalase. Thus, we sought to immunohistochemically determine the endogenous presence of the H₂O₂-reducing antioxidant enzyme catalase within nTS neurons. Specificity of the antibody for catalase in nTS tissue was verified, in part, via western blot analysis (Figure 8A). A single band for catalase was found as expected at ~65 kDa. Strong catalase-IR was found ubiquitously throughout the whole nTS with a few scattered cells exhibiting very strong IR for catalase (Figure 8B). In NeuN-identified neurons, catalase-IR was primarily localized surrounding DAPI-labeled nuclei (perinuclear) to a dense network that extended into the cellular processes (Figure 8C).

Figure 8. Distribution of catalase within the nTS.

A. Analysis of antibody specificity via immunoblots for 20 μg of protein from the nTS. A single band for catalase was confirmed at the appropriate size of ~65kDa. B Catalase-IR (pseudocolored green) in a coronal brainstem section. Catalase-IR is present ubiquitously throughout the nTS (dashed line). Spinning disc confocal image. AP = area postrema, TS = tractus solitarii, CC = central canal. Scale bar, 200 μm. C Higher magnification of catalase (green) with co-labeling of neuronal cells (NeuN, blue) and nuclei (DAPI, grey). Merged image is shown in the bottom right panel. Representative images are a maximal projection of a z-stack with 20 images 0.5 μm apart. Catalase-IR was identified perinuclear in NeuN-identified neurons. Scale bar, 15 μm.

DISCUSSION

Our data demonstrate H₂O₂ alters intrinsic and firing properties of nTS neurons while not affecting synaptic transmission from visceral afferents. The H₂O₂-induced immediate postsynaptic changes included hyperpolarization of RMP and THR, increase in rheobase, and reduction of AP discharge. H₂O₂ increased slope conductance of barium-sensitive IRK and/or K_2P channels, and Ba²⁺ blocked the H₂O₂ changes in RMP, R_i and AP discharge. Intracellular application of the antioxidant enzyme catalase, which is ubiquitously expressed in nTS neurons, prevented H₂O₂-mediated effects on RMP and THR. The H₂O₂-effect on RMP, but not THR, was fully reversible and produced hyperexcitability in response to greater depolarization. Altogether, these data suggest H₂O₂ is an important modulator of neuronal activity in the nTS.

In the present study, we used short exposures of H₂O₂ at concentrations to approximate levels typically observed in vivo. Specifically, 0.01 – 1 mM exogenous H₂O₂ in the central nervous system has been shown to mimic the concentration of H₂O₂ released endogenously by cellular effectors [e.g., NADPH oxidase, mitochondria, and xanthine oxidase (Schroder and Eaton, 2008)], due in part to the 10-fold reduction of H₂O₂ concentration by spontaneous and antioxidant-mediated breakdown. Given these assumptions, the intracellular H₂O₂ concentrations achieved in our study are likely to range from 1 – 50 μM. Furthermore, our concentrations are considerably less than previous studies which have illustrated an effect of millimolar concentrations of H₂O₂ on synaptic or neuronal function (Frantseva et al., 1998; Avshalumov et al., 2005; Pouokam et al., 2009; Garcia et al., 2011). H₂O₂ concentrations greater than 10 μM and up to 1 mM have been shown to occur physiologically but are also associated with oxidative stress (Stone and Yang, 2006; Schroder and Eaton, 2008). Thus, our results may be directly correlated to a variety of conditions and provide mechanistic insight into the role of H₂O₂ in the nTS on the cardiorespiratory system.

Previous studies using similar concentrations to ours (up to 500 μM for ~30 minutes) demonstrated that H₂O₂ alters synaptic transmission. For instance, H₂O₂ reduces hippocampal long-term potentiation (Auerbach and Segal, 1997) and EPSP amplitude (Kamsler and Segal, 2003), and increases GABA release (Takahashi et al., 2007). In the current experiments, H₂O₂ did not alter postsynaptic CNQX-sensitive non-NMDA glutamate receptor properties at the TS-nTS synapse, as indicated by the lack of change in evoked and miniature EPSC amplitude, and TS-EPSC decay time. Likewise, the lack of effect on PPR, use-dependent depression, as well as asynchronous and miniature EPSC frequency suggests that H₂O₂ did not alter glutamate release from TS terminals. Across all cells tested, H₂O₂ also did not alter TS-EPSCs regardless of initial current amplitude or failure rate (not shown), indicators of synapses containing myelinated (A-type) or unmyelinated (C-type) afferents (Andresen and Peters, 2008), suggesting the H₂O₂ response was not fiber type specific. Conversely, 500 μM H₂O₂ decreased sEPSC frequency, but not amplitude, in the neurons studied. Given that sEPSCs derive from network activity within the available nTS circuitry (Fortin and Champagnat, 1993), the decrease in sEPSC frequency may indicate that H₂O₂ reduced the activity of neurons which form a synapse with the recorded cell. This is further supported by the lack of change in mEPSC frequency during H₂O₂.

In the present study, AP discharge decreased during the initial bout of H₂O₂. This likely resulted from alteration in postsynaptic cell properties independent of glutamate receptors, including R_i, RMP and THR. Specifically, H₂O₂ hyperpolarized RMP and THR, and reduced R_i. These effects are likely inter-related. The H₂O₂-induced decrease in R_i suggests opening, or increased conductance, of membrane channel(s) that contribute to RMP hyperpolarization and the decrease in discharge, as in cells from other central nuclei (Seutin et al., 1995; Avshalumov et al., 2005; Pouokam et al., 2009; Garcia et al., 2011). Overall, changes in R_i, RMP and THR are also responsible for the increase in rheobase during peroxide exposure. H₂O₂ hyperpolarized RMP toward our calculated K⁺ equilibrium potential, suggesting an involvement of K⁺-channels. H₂O₂ activates ATP-sensitive K⁺-channels (Avshalumov et al., 2005), an inward rectifying K⁺-channel that modulates RMP in nTS neurons during hypoxia (Duprat et al., 2005; Kim et al., 2007; Zhang et al., 2008). Additional IRK and K_2P channels have also been demonstrated within the nTS (Dunn-Meynell et al., 1998; Gabriel et al., 2002; Yamamoto et al., 2008). To identify the effect of H₂O₂ on IRK and K_2P channels in nTS neurons, we recorded holding currents in H₂O₂-responsive cells during depolarizing ramps. Results demonstrated that H₂O₂ augmented slope conductance, which was blocked or prevented by barium at a concentration specific for IRK/K_2P currents (Bayliss et al., 1997; Day et al., 2005; Austgen et al., 2012). Since the augmented H₂O₂-sensitive current reversed towards the calculated K⁺ equilibrium potential and was blocked by barium, these data suggest that H₂O₂ augments IRK/K_2P conductance. The holding current during H₂O₂ had an overall linear shape suggesting H₂O₂ has a main effect on K_2P conductance, although alterations of IRK channel conductance cannot be excluded. Also, the positive shift in reversal potential may suggest that additional channels contribute. For instance, H₂O₂ also increases conductance of non-selective cation channels (Simon et al., 2002; Sato et al., 2009) and sodium channels (Meng and Nie, 2004). The potential increase of sodium influx by H₂O₂ through these channels is consistent with greater ease to reach THR in our current study, as demonstrated in our DE protocols. Regardless, the decrease in discharge during the initial bout of H₂O₂ is likely due to the reduction in R_i and the concurrent change in RMP, reducing the stimulus effect and increasing rheobase. Importantly, action potential discharge was reduced by H₂O₂ in response to current steps in the range of our TS-EPSC amplitudes. Thus it is likely that sensory afferent or network evoked currents would also be less effective in the presence of H₂O₂.

We observed that following washout of the initial bout of H₂O₂, AP discharge increased above baseline, especially during greater depolarizing currents. This was observed whether AP discharge was monitored via step or ramp depolarization. This is likely due to the fact that during wash RMP readily returned to baseline with an increase of R_i (likely due to closing of K⁺ channels), whereas THR remained hyperpolarized in comparison to baseline. This shift significantly reduced [THR-RMP] and returned rheobase to or below baseline values. Therefore, hyperexcitability and an enhancement of AP discharge were observed in response to a depolarizing stimulus. It is anticipated that network or afferent-driven increases in activity would also produce elevated AP discharge. R_i partially reversed, consistent with the closing of H₂O₂-sensitive K⁺ channels and the return of RMP to baseline. Because R_i does not fully reverse, it is possible that other channels, which may contribute to changes in THR, remain in a state of increased conductance. This would be consistent with a persistent hyperpolarization of THR. One should note that the majority of H₂O₂-inhibitory effects occurred primarily in the initial responses, with subsequent exposures being less effective. Thus, AP discharge was significantly greater during the second H₂O₂ exposure compared to the first exposure and to baseline. The persistent increase in discharge may be due to protein modification of the ion channel(s) altered. This notion however requires further study. Of interest, sustained hyperexcitability with H₂O₂ (subsequent to a transient decrease of excitability) has also been reported in neurons of the pre-Bötzinger complex (Garcia et al., 2011).

We also demonstrate that H₂O₂ modulates AP firing in DE neurons. The typical delay of discharge following a hyperpolarizing pre-pulse, a phenomenon attributed to the relief of A-type K⁺ channel inactivation (Schild et al., 1993), was observed during baseline and H₂O₂. H₂O₂ decreased the overall delay of depolarization-induced firing (i.e., neurons fired earlier) across the hyperpolarizing pre-pulses, yet the slope of the relationship between pre-pulse potential and magnitude of delay was unchanged. Such changes suggest the magnitude of A-type channel currents are not altered by H₂O₂. Furthermore, even though there was a tendency to decrease A-type currents during H₂O₂, which would promote a decrease in delay and increase in firing, it was not significant in the 9 cells studied. However, we did not examine whether transient A-type channel kinetics were altered, including inactivation rate and voltage-dependence of activation and inactivation, which may substantially alter firing as they do in other central neurons, including the PVN (Sonner and Stern, 2007). Future studies in isolated nTS cells which lack many of the space constant and pharmacological constraints inherent in the slice preparation will further elucidate if H₂O₂ alters channel kinetics. Preventing the H₂O₂-induced hyperpolarization of RMP with bias current increased AP discharge and decreased its delay in the majority of cells, due to a decrease in THR. Taken together, these results further suggest that the initial decrease in AP discharge by H₂O₂ is likely due to membrane hyperpolarization (and the interrelated increase in membrane conductance), and that the reduction in THR coupled with the depolarized RMP is responsible for the delayed hyperexcitability seen following washout.

H₂O₂ is endogenously eliminated by intracellular antioxidants, including catalase. We observed catalase-IR throughout the nTS and dorsal motor complex. Catalase-IR was primarily perinuclear and organized to a dense structure in nTS neurons. Catalase has been located in peroxisomes (Mavelli et al., 1982) and mitochondria (Radi et al., 1991), which can be closely associated with cisternae of the endoplasmic reticulum (Schrader and Fahimi, 2004). The importance of intracellular catalase in breaking down H₂O₂ is evidenced in our protocols in which we elevated intracellular catalase via its placement in the recording pipette. We reasoned that if H₂O₂ modulated RMP and THR via an intracellular mechanism, then supplemental catalase should reduce or eliminate its effects. Our results confirm this notion and indicate that H₂O₂ likely alters postsynaptic activity via intracellular rather than extracellular sites.

H₂O₂ may be converted to hydroxyl radical (·OH), which itself may alter neuronal activity (Avshalumov et al., 2000; Garcia et al., 2011). While we cannot completely rule out that H₂O₂ exerts some of its action through ·OH, the ablation of H₂O₂-induced changes on RMP and THR by supplemental catalase suggests that the contribution of ·OH may be small. Particularly, if ·OH mediates part or all of the actions of H₂O₂, then ·OH produced by exogenous H₂O₂, either spontaneously in the bath or in nearby cells, would likely continue to alter RMP and THR even during elevated catalase. This did not occur. While the availability of free iron, which is needed to produce ·OH, in an isolated nTS slice is unknown, it is reasonable to suggest it may be limited, especially during repetitive cycles of H₂O₂ perfusion and washout. Thus, the observed reduction of cell excitability in nTS cells in the current experiments is likely mediated by H₂O₂.

In summary, we demonstrate that H₂O₂ modulates nTS neuronal function. These studies provide insight into the functional consequences of H₂O₂ under normal conditions and disease states. The medial and commissural nTS are the primary afferent termination sites for multiple visceral reflexes, including the gastrointestinal, cardiovascular and respiratory systems (Andresen and Kunze, 1994; Travagli et al., 2006; Dean and Putnam, 2010). H₂O₂ in the nTS may profoundly alter these reflexes individually as well as their coordination (Dean, 2011). While the effects of nTS H₂O₂ have not been examined for the gastrointestinal system, one can speculate that the prolonged hyperexcitability in the presence of H₂O₂ may augment the inhibition of gastric tone or motility that occurs through the vaso-vagal reflex (Travagli et al., 2006). H₂O₂ may also alter the coordination between the respiratory and digestive system for pH regulation that has been proposed to occur in the nTS (Dean, 2011). In regards to the cardiovascular system, microinjection of H₂O₂ into the nTS evokes hypotension and bradycardia (Cardoso et al., 2009), consistent with an excitatory role on cells in the baroreflex pathway. Likewise, ROS have been implicated in the disease manifestation following chronic intermittent hypoxia (CIH), an animal model for obstructive sleep apnea (OSA) (Sica et al., 2000). Both CIH and OSA cause increased chemoreflex sensitivity, augmented sympathetic nerve activity and hypertension (Fletcher et al., 1992; Somers and Abboud, 1993; Prabhakar et al., 2011). While an increase in peripheral chemoreceptor activity contributes to CIH-induced sympathoexcitation (Peng and Prabhakar, 2004; Peng et al., 2004), a substantial contribution is also ascribed to the central component (Ling et al., 2001; Kline et al., 2007). ROS are strongly implicated to mediate these changes in the cardiorespiratory system (Kuo et al., 2011; Prabhakar, 2011) when produced in excess during bouts of hypoxia-reoxygenation (Prabhakar, 2001). Thus, increased neuronal excitability due to H₂O₂ within the reflex neuroaxis (including arterial chemoreceptors, their afferents and central terminals, interneurons and projection nTS neurons, as well as other important cardiorespiratory nuclei) may occur during each hypoxic bout and account for increasingly augmented sympathetic nerve activity observed in CIH or OSA. Our results demonstrate a biphasic response to exogenous H₂O₂ that ultimately leads to cell hyperexcitability, which may contribute to elevated sympathetic nerve activity through downstream nTS projections. Moreover, given the potential therapeutic use of antioxidants, these results may provide insight into the mechanism and their site of action.

Highlights.

• We studied effects of H₂O₂ on synaptic and intrinsic properties of nTS neurons.

• Catalase blocked effects of H₂O₂ to decrease membrane potential and AP threshold.

• Conductance through IRK/K_2P channels was increased with H₂O₂.

• H₂O₂ initially depressed discharge, yet resulted in sustained hyperexcitability.

• The anti-oxidant enzyme catalase is present throughout the nTS.

Glossary

aEPSC

asynchronous EPSC

AP

action potential

DE

delayed excitation

H₂O₂

hydrogen peroxide

IR

immunoreactivity

IRK

inward rectifying K⁺ channel

mEPSC

miniature EPSC

nTS

nucleus tractus solitarii

O₂^·−

superoxide anion

·OH

hydroxyl radical

K_2P

tandem-pore domain K⁺ channel

PPR

paired pulse ratio

PSC

postsynaptic current

R_i

input resistance

RMP

resting membrane potential

ROS

reactive oxygen species

sEPSC

spontaneous EPSC

THR

action potential threshold

TS

tractus solitarii

TS-EPSCs

evoked EPSCs