Abstract
Nonselective cation channels promote persistent spiking in many neurons from a diversity of animals. In the hermaphroditic marine-snail, Aplysia californica, synaptic input to the neuroendocrine bag cell neurons triggers various cation channels, causing an ∼30 min afterdischarge of action potentials and the secretion of egg-laying hormone. During the afterdischarge, protein kinase C is also activated, which in turn elevates hydrogen peroxide (H2O2), likely by stimulating nicotinamide adenine dinucleotide phosphate oxidase. The present study investigated whether H2O2 regulates cation channels to drive the afterdischarge. In single, cultured bag cell neurons, H2O2 elicited a prolonged, concentration- and voltage-dependent inward current, associated with an increase in membrane conductance and a reversal potential of ∼+30 mV. Compared with normal saline, the presence of Ca2+-free, Na+-free, or Na+/Ca2+-free extracellular saline, lowered the current amplitude and left-shifted the reversal potential, consistent with a nonselective cationic conductance. Preventing H2O2 reduction with the glutathione peroxidase inhibitor, mercaptosuccinate, enhanced the H2O2-induced current, while boosting glutathione production with its precursor, N-acetylcysteine, or adding the reducing agent, dithiothreitol, lessened the response. Moreover, the current generated by the alkylating agent, N-ethylmaleimide, occluded the effect of H2O2. The H2O2-induced current was inhibited by tetrodotoxin as well as the cation channel blockers, 9-phenanthrol and clotrimazole. In current-clamp, H2O2 stimulated burst firing, but this was attenuated or prevented altogether by the channel blockers. Finally, H2O2 evoked an afterdischarge from whole bag cell neuron clusters recorded ex vivo by sharp-electrode. H2O2 may regulate a cation channel to influence long-term changes in activity and ultimately reproduction.
SIGNIFICANCE STATEMENT Hydrogen peroxide (H2O2) is often studied in a pathological context, such as ischemia or inflammation. However, H2O2 also physiologically modulates synaptic transmission and gates certain transient receptor potential channels. That stated, the effect of H2O2 on neuronal excitability remains less well defined. Here, we examine how H2O2 influences Aplysia bag cell neurons, which elicit ovulation by releasing hormones during an afterdischarge. These neuroendocrine cells are uniquely identifiable and amenable to recording as individual cultured neurons or a cluster from the nervous system. In both culture and the cluster, H2O2 evokes prolonged, afterdischarge-like bursting by gating a nonselective voltage-dependent cationic current. Thus, H2O2, which is generated in response to afterdischarge-associated second messengers, may prompt the firing necessary for hormone secretion and procreation.
Introduction
Reactive oxygen species are generated either as a byproduct of oxidative metabolism (Babior et al., 1973; Kourie, 1998) or via NADPH oxidase (Bedard and Krause, 2007). O2 is reduced to the superoxide anion radical, O2•−, which is in turn dismutated to the more stable and freely diffusible hydrogen peroxide (H2O2; Brookes et al., 2004; Bedard and Krause, 2007; Lee et al., 2015a). Elevated reactive oxygen species can lead to stress and cell death, potentially contributing to aging (Sohal and Orr, 2012) or Alzheimer's and Parkinson's diseases (Hernandes and Britto, 2012). However, growing evidence shows that H2O2 may regulate ion channels physiologically; for example, in the substantia nigra, H2O2 hyperpolarizes dopaminergic neurons by gating an ATP-sensitive K+ channel, and depolarizes GABAergic neurons by gating a nonselective cation channel (Avshalumov et al., 2005; Lee et al., 2011).
Cation channels underlie bursting and persistent firing in many neurons (Partridge et al., 1994; Major and Tank, 2004). This includes cation channel-dependent protracted spiking in the snails, Aplysia (Kramer and Zucker, 1985; Matsumoto et al., 1988), Archidoris (Partridge et al., 1979), and Helix (Swandulla and Lux, 1985; Partridge and Swandulla, 1987). For rodents, cation channels mediate long-lasting activity in lumbosacral spinal cord (Derjean et al., 2005), dorsal horn (Morisset and Nagy, 2000), olfactory bulb (Shpak et al., 2012), substantia nigra (Mrejeru et al., 2011), and rostral ambiguus (Rekling and Feldman, 1997), as well as hippocampal (Knauer et al., 2013), anterior cingulate (Ratté et al., 2018), entorhinal (Tahvildari et al., 2008), and prefrontal cortex (Yan et al., 2009; Baker et al., 2018). Functionally, cation channels impact memory (Egorov et al., 2002; Sidiropoulou et al., 2009), sensory coding (Dong et al., 2009), motor pattern generation (Di Prisco et al., 1997), and neuroendocrine control (Chakfe and Bourque, 2000).
The bag cell neurons are neuroendocrine cells that initiate reproduction in Aplysia (Conn and Kaczmarek, 1989; Zhang and Kaczmarek, 2008; Sturgeon et al., 2018). In response to cholinergic input, these neurons depolarize and undergo a lengthy afterdischarge, with both a fast-phase (∼5 Hz, ∼1 min) and slow-phase (1 Hz, ∼30 min) of action potential firing that results in the neurohemal secretion of egg-laying hormone (Kupfermann and Kandel, 1970; Arch, 1972; Pinsker and Dudek, 1977; Loechner et al., 1990; Roubos et al., 1990; Michel and Wayne, 2002; Hatcher and Sweedler, 2008; White and Magoski, 2012). The hormone engages both central neurons and peripheral organs to induce behaviors that culminate in the deposition of fertilized eggs (Arch and Smock, 1977; Stuart and Strumwasser, 1980; Rothman et al., 1983). The afterdischarge is maintained by opening of three, distinct cation channels: a voltage-independent channel gated by calmodulin-kinase (Hung and Magoski, 2007; Hickey et al., 2010), a separate voltage-independent channel triggered by diacylglycerol (DAG; Sturgeon and Magoski, 2016), and a Ca2+-permeable, Ca2+-activated, voltage-dependent channel (Wilson et al., 1996; Lupinsky and Magoski, 2006; Geiger et al., 2009). Munnamalai et al. (2014) showed that NADPH oxidase is present in bag cell neurons, and protein kinase C (PKC), which is activated early in the slow-phase of the afterdischarge (Wayne et al., 1999), stimulates H2O2 production. This likely occurs through PKC-mediated phosphorylation of the NADPH oxidase cytosolic regulatory subunit, p47phox (Fontayne et al., 2002). Thus, we sought to test whether H2O2 can affect bag cell neuron function in a manner concordant with afterdischarge generation. The present study shows that exogenous H2O2 triggers a conductance similar to the voltage-dependent cation channel previously characterized in bag cell neurons. This H2O2-induced current is sensitive to both changes in redox and cation channel blockers; moreover, it produces prolonged depolarization and firing. Historically, excessive H2O2 has been implicated in neuronal cell death (Herson and Ashford, 1997; Herson et al., 1999; Smith et al., 2003); however, along with work on substantia nigra (Lee et al., 2011, 2013) and hippocampal (Olah et al., 2009) neurons, our results suggest a role for H2O2 in regulating persistent firing. In Aplysia, such regulation may have consequences for procreation.
Materials and Methods
Animals and cell culture.
Adult Aplysia californica (a hermaphrodite) weighing 200–650 g were obtained from Marinus and housed in an ∼300 L aquarium containing continuously circulating, aerated sea water (Instant Ocean, Aquarium Systems) at 16–18°C on a 12:12 h light/dark cycle and fed romaine lettuce five times/week. All experiments were approved by the Queen's University Animal Care Committee (protocols 2013-041 and 2017-1745).
For primary cultures of isolated bag cell neurons, animals were anesthetized by an injection of isotonic MgCl2 (0.39 m; volume ∼50% of body weight), and the abdominal ganglion was removed and treated with dispase II (13.3 mg/ml; 04942078601, Roche Diagnostics/Sigma-Aldrich) dissolved in tissue culture artificial sea water (tcASW; composition in mm: 460 NaCl, 10.4 KCl, 11 CaCl2, 55 MgCl2, 15 HEPES, 1 mg/ml glucose, 100 U/ml penicillin plus 0.1 mg/ml streptomycin (P4333, Sigma-Aldrich), pH 7.8 with NaOH, for 18 h at 22°C. The ganglion was then rinsed in tcASW for 1 h, and the two bag cell neuron clusters were dissected from their surrounding connective tissue. Using a fire-polished glass Pasteur pipette and gentle trituration, neurons were dissociated and dispersed in tcASW onto 35 × 10 mm polystyrene tissue culture dishes (353001, Falcon-Corning/ThermoFisher Scientific). Cultures were maintained in a 14°C incubator and used for experimentation within 1–3 d. Salts were obtained from ThermoFisher Scientific, MP Biomedicals, Acros Organics, or Sigma-Aldrich.
Whole-cell voltage-clamp and current-clamp recording from cultured bag cell neurons.
Tight-seal, whole-cell recordings of membrane current or voltage from cultured bag cell neurons were performed at room temperature (20–22°C) using an EPC-8 amplifier (HEKA Electronik/Harvard Apparatus). Borosilicate-glass microelectrodes had a resistance of 1–2 MΩ when filled with various intracellular saline (see following paragraph). Pipette and membrane capacitive currents were cancelled and the series resistance (3–5 MΩ) compensated to 70–80%. Current was low-pass filtered at 1 kHz and sampled at 2 kHz using a Digidata 1322A analog-to-digital converter (Molecular Devices) and the Clampex acquisition program of pCLAMP v8.1 (Molecular Devices). Holding and test potentials (see Results) were delivered using pCLAMP. Voltage was low-pass filtered at 5 kHz and sampled as per current; in addition, membrane potential was initially set to −60 or −40 mV with constant bias current from the EPC-8.
Unless otherwise noted, most recordings were made in normal artificial sea water (nASW; composition as per tcASW but lacking glucose and antibiotics) with a Cs+-based intracellular saline (composition in mm: 500 Cs+-Asp, 70 KCl, 1.25 MgCl2, 10 HEPES, 11 glucose, 10 glutathione, 5 EGTA, 5 adenosine 5′-triphosphate 2Na·H2O (A3377, Sigma-Aldrich), and 0.1 guanosine 5′-triphosphate Na·H2O (G8877, Sigma-Aldrich), pH 7.3 with KOH. Some experiments used a K+-based intracellular saline as per the Cs+-based saline, but with K+ replacing Cs+. For both intracellular salines, as calculated using WebMaxC (https://somapp.ucdmc.ucdavis.edu/pharmacology/bers/maxchelator/webmaxc/webmaxcS.htm), 3.75 mm CaCl2 was added to set the free Ca2+ concentration at 300 nm, which corresponds to the approximate resting intracellular Ca2+ concentration of bag cell neurons as determined using either Ca2+-sensitive electrodes (Fisher et al., 1994) or imaging of Ca2+-sensitive dyes (Fink et al., 1988; Loechner et al., 1992; Knox et al., 1996; Magoski et al., 2000). Junction potentials of 17 and 15 mV were calculated for the Cs+-based and K+-based intracellular salines, respectively, versus nASW and compensated for by subtraction off-line. In a small number of experiments, voltage-gated Ca2+ current was recorded by using the Cs+-based intracellular saline and an ASW where all of the Na+ and K+ were replaced with tetraethylammonium and Cs+, respectively. The 20 mV junction potential for this saline combination was compensated for by subtraction off-line.
Experiments designed to examine the reversal potentials of the H2O2-induced current involved external Na+ and Ca2+ substitutions to make Na+-free [composition as per nASW but with the Na+ replaced by N-methyl-d-glucamine (NMDG)], Ca2+-free (Ca2+ replaced by Mg2+), and Na+/Ca2+-free (Na+ replaced by NMDG and Ca2+ replaced by Mg2+) salines. Junction potentials of 17 mV for Ca2+-free ASW and 23 mV for both Na+-free ASW and Na+/Ca2+-free ASW versus the Cs+-based intracellular saline were compensated for by subtraction off-line.
Ensemble extracellular recording from the intact bag cell neuron cluster.
For extracellular recording, the abdominal ganglion, including the two bag cell neuron clusters and associated pleuroabdominal connectives, was isolated and maintained in nASW-filled dish kept at a 14°C by immersion in a water-cooled chamber. A wide-bore, fire-polished glass suction recording electrode (containing nASW) was placed over one of the two bag cell neuron clusters, while a similarly fashioned stimulating electrode was placed at the rostral end of the pleuroabdominal connective ipsilateral to the recorded cluster. Synaptic input was stimulated with current from a Grass SD9 stimulator (Astro-Med) while voltage was monitored with a Model 3000 AC/DC differential amplifier (A-M Systems). Voltage was high-pass filtered at 10 Hz and low-pass filtered at 1 kHz, and acquired at 2 kHz using AxoScope v9.0 (Molecular Devices) as per whole-cell voltage-clamp.
Sharp-electrode recording from bag cell neurons in culture and isolated clusters.
Sharp-electrode recordings from bag cell neurons were performed using an AxoClamp 2B amplifier (Molecular Devices) at room temperature (20–22°C). Borosilicate-glass microelectrodes had a resistance of 5–20 MΩ when filled with 2 m K+-acetate plus 10 mm HEPES and 100 mm KCl, pH 7.3 with KOH. Voltage-clamp was performed on single, cultured bag cell neurons using continuous single-electrode voltage-clamp at a holding potential of −30 mV. Whereas current-clamp was undertaken on bag cell neurons in desheathed, isolated clusters using the bridge-balance method. A Grass S88 stimulator was used to deliver 50 ms hyperpolarizing current pulses to balance the bridge; in addition, bias current was injected into the neuron as necessary from the AxoClamp. Current or voltage was filtered to 3 kHz before sampling at 2 kHz as per whole-cell voltage-clamp with Clampex.
Drug application and reagents.
Solution exchanges were initially accomplished using a calibrated transfer pipette to replace the bath (culture dish) tcASW with the desired extracellular saline before the start of a given experiment. In most cases, drugs were introduced before or during recording by initially removing a small volume (∼50 μl) of saline from the bath, combining that with an even smaller volume (<10 μl) of drug stock-solution (see following paragraph), and then reintroducing that mixture back into the bath. Care was taken to pipette near the side of the dish and as far away as possible from the neuron. For some experiments, drugs or transmitters were applied using a single-cell superfusion system consisting of a micromanipulator-controlled square-barreled glass pipette (∼500 μm bore; 8250, VitroCom) positioned 300–500 μm from the soma and connected by a stopcock manifold to a series of gravity-driven reservoirs. This provided a constant flow (∼0.5–1 ml/min) of control extracellular saline over the neuron, which was switched to drug-containing saline by activating the appropriate stopcock. The bath volume was kept constant during superfusion with vacuum-trap suction outlet.
Water was used to dissolve the following as stocks at the indicated concentrations: H2O2 0.2 m (H325–500, ThermoFisher Scientific), mercaptosuccinic acid (0.3 m; M6182; Sigma-Aldrich), acetylcholine chloride (ACh; 100 mm; A6625, Sigma-Aldrich), DL-dithiothreitol (DTT; 50 mm; 646563, Sigma-Aldrich), 1-(2-(3-(4-methoxyphenyl)propoxy)-4-methoxyphenylethyl)-1H-imidazole (SKF-96365; 50 mm; S7809, Sigma-Aldrich), tetrodotoxin (TTX; 3 mm; T-550, Alomone Labs), and N-acetyl-l-cysteine (NAC; 150 mm; A7250, Sigma-Aldrich). Similarly, dimethyl sulfoxide (DMSO; BP231-1, ThermoFisher Scientific) was used to dissolve the following: clotrimazole (25 mm; C6019, Sigma-Aldrich), and 9-phenanthrol (9-Pt; 50 mm; 211281, Sigma-Aldrich). Last, ethanol (100% v/v) was used to dissolve N-ethylmaleimide (NEM; 50 mm; E3876, Sigma-Aldrich). The final concentration of DMSO or ethanol in the bath was ≤0.2% (v/v), which in control experiments here or in prior studies was found to have no effect on bag cell neuron holding current, membrane conductance, or membrane potential (Hickey et al., 2010, 2013; Tam et al., 2011; Sturgeon and Magoski, 2016; White et al., 2018).
Analysis.
The Clampfit analysis program of pCLAMP was used to determine the amplitude of changes to membrane current or potential evoked by H2O2 or other reagents under voltage- or current-clamp. For peak change, two cursors were placed 30 s apart, 30 s before drug addition, the average between the cursors served as a baseline. An additional two cursors were placed 60 s apart on either side of the peak of the response. Clampfit then calculated the peak amplitude relative to the baseline. Percent recovery of the current following a response was determined by comparing the peak current to the steady-state current, the latter being calculated by again placing two cursors, 30 s apart, at the end of the trace, well after the maximal response and where the response had recovered to steady-state. For display only, some current traces were filtered off-line to between 50 and 100 Hz using the Clampfit Gaussian filter. Due to the overall slow nature of the responses, this second filtering brought about no change in amplitude or kinetics. Conductance was derived using Ohm's law (G = I/V) and the current change during a 200 ms step from −30 to −40 mV before and after H2O2. Reversal potential involved taking a difference current, ascertained by subtracting the current elicited by a voltage ramp from −60 to +60 mV before H2O2 superfusion, from the current elicited by the same ramp at the peak of the H2O2-induced current. The reversal potential was then measured directly from the difference current by placing a cursor where the current crossed the y-axis. For some responses recorded under current-clamp that presented with robust spiking, which made it difficult to visualize the peak depolarization, Clampfit was used to generate all-point histograms for before and after H2O2 application. The largest peaks of the resulting histograms were fit with a Gaussian function by the least-squares method and a simplex search, and taken as the average membrane potential. The difference between the voltage before and after H2O2 served as the amplitude of the depolarization. To determine the frequency of firing caused by H2O2, the number of spikes during the response was determined using the Clampfit threshold search function, and was divided by the total time of the burst (in seconds) to derive action potential frequency.
Experimental design and statistical analysis.
Data are mean ± SEM. Statistical analysis was performed using InStat v3.10 and Prism v8.0.0. The Kolmogorov–Smirnov method was used to test for normality. To test whether the mean differed between two groups of normally distributed data, Student's unpaired t test with Welch correction as necessary was used, whereas for not normally distributed data, the Mann–Whitney U test or Wilcoxon matched-pairs signed ranks test was used. For three or more means, normally distributed data were compared using an ordinary one-way ANOVA followed by the Tukey-Kramer multiple-comparisons test or Dunnett multiple-comparisons test, whereas not normally distributed data were compared using a Kruskal–Wallis ANOVA (KW-ANOVA) followed by Dunn's multiple-comparisons test. Significance level was a two-tailed p value of < 0.05. Prism was also used to fit the H2O2 concentration–response relationship with a four-parameter dose–response equation to determine the half-maximal concentration (EC50) and the Hill slope (steepness of the curve).
Results
H2O2 activates a prolonged, inward voltage-dependent current in cultured bag cell neurons
H2O2 has been shown to induce changes in the excitability of sensory neurons in Aplysia, as well as striatal, substantia nigra, spinal ventral horn, hippocampal, or cortical neurons (Chen et al., 2001; Chang et al., 2003; Olah et al., 2009; Lee et al., 2011, 2013; Ohashi et al., 2016). Given that PKC, which is upregulated 2–5 min from the onset of the afterdischarge (Wayne et al., 1999), can increase H2O2 production in bag cell neurons (Munnamalai et al., 2014), we examined the effects of extracellularly applied H2O2 on bag cell neurons in primary culture.
Initially, the efficacy of the superfusion apparatus was assessed to ascertain the exact time of H2O2 delivery to the neuron. This involved superfusing (see Materials and Methods, Drug application and reagents) 1 mm ACh over the soma of a neuron under whole-cell voltage-clamp at −60 mV (Fig. 1A, left). We previously established that ACh gates, with essentially no delay, an ionotropic receptor in bag cell neurons (White and Magoski, 2012). This ligand-gated channel opens as soon as ACh contacts the cell, and is sensitive to classic nicotinic-type agonists and antagonists (White and Magoski, 2012; White et al., 2014). Thus, the latency from when the ACh-containing reservoir was opened to when the ACh-induced current was first observed represented the lag time of our perfusion apparatus.
H2O2 activates a prolonged, inward, voltage-dependent current in cultured bag cell neurons. A, Left, Phase-contrast photomicrograph showing a superfusion barrel positioned near a bag cell neuron soma with neurites and the recording pipette. Neurons are whole-cell voltage-clamped at −60 mV using our standard K+-Asp-based intracellular saline in the pipette and nASW in the bath. Middle, Superfusion of 1 mm ACh (at bar) elicits an inward current. Arrows highlight the start of superfusion and travel time, i.e., the time required for ACh to reach the soma. Right, ACh superfusion produced a peak current (IACh) of −2.9 ± 1.6 nA with a latency of 38.0 ± 5.3 s following the switch from control to ACh-containing saline. B, Superfusion of 1 mm H2O2 (at bar) over the soma of different neurons, held at −60, −30, or 0 mV, causes increasingly larger inward current. Ordinate applies to all traces. C, Summary current/voltage relationship for the H2O2-induced current at −60 mV (91.0 ± 43 pA), −40 mV (51.0 ± 31.0 pA), −30 mV (230.0 ± 74.0 pA), −20 mV (288.0 ± 84.0 pA), or 0 mV (765.0 ± 138.0 pA), shows a threshold between −40 and −30 mV and an apparent maximum at 0 mV. D, Group data illustrating a 59.9 ± 12.4 s latency of the H2O2-induced current at −30 mV.
Neurons were bathed in Na+-containing nASW and dialyzed for 15 min with our standard K+-Asp-based intracellular saline (see Materials and Methods, Whole-cell voltage-clamp and current-clamp recording from cultured bag cell neurons). ACh induced an inward current of ∼3 nA with a mean onset latency of 38 ± 5.3 s (n = 4) from the time superfusion began, i.e., when the stopcock on the reservoir was switched (Fig. 1A, middle, right). Subsequently, it was assumed that when H2O2 was superfused it took 38 s to reach the soma. Parenthetically, we performed a second calibration using block of voltage-gated Ca2+ current by 10 mm Ni2+, which inhibits the Ca2+ current almost instantly (Hung and Magoski, 2007). Ca2+ current was evoked at a frequency of 1 Hz using a 75 ms step from −60 to 0 mV; when the perfusion was switched from control saline to Ni2+-containing saline, it required 40 ± 3.5 s for Ni2+ to initiate block (n = 10). This lag time was not significantly different from that observed for response onset of the ACh-induced current (t(12) = 0.3097; p = 0.7621, unpaired Student's t test).
Superfusion of 1 mm H2O2 dissolved in nASW while holding the membrane potential at −60, −40, −30, −20, or 0 mV elicited a voltage-dependent inward current (Fig. 1B). There was minimal current (<100 pA) at −60 mV (n = 6) or −40 mV (n = 4), whereas more moderate currents (200–300 pA) were evoked at −30 (n = 4) or −20 mV (n = 6), with the maximal current (∼800 pA) at 0 mV (n = 5; Fig. 1C). At −30 mV, the latency from when H2O2 arrived at the cell to when the current began was ∼60 s (n = 4; Fig. 1D).
A concentration-dependent H2O2-induced inward current
To assess the sensitivity of bag cell neurons to H2O2, we examined the effect of various concentrations. Previous reports, using both vertebrate and invertebrate neurons, used a range of H2O2 from 1 to 3 mm, which did not cause damage in the short-term (Chen et al., 2001; Chang et al., 2003; Olah et al., 2009; Lee et al., 2011, 2013; Ohashi et al., 2016). Additionally, limitations on membrane permeability and intracellular scavenging suggests that the H2O2 concentration inside a cell is at least tenfold (perhaps even 650-fold) less than the concentration of extracellularly applied H2O2 (Antunes et al., 2000; Miller et al., 2007; Huang and Sikes, 2014).
Because the H2O2-induced current was rather small at −60 mV, and the holding current was often not entirely stable at 0 mV, we chose to examine the response at −30 mV in subsequent recordings. This provided a more uniform H2O2-induced current with a reasonably stable baseline. Cultured bag cell neurons were whole-cell voltage-clamped in nASW using a Cs+-Asp-based intracellular saline to block K+ channels and further improve resolution at this somewhat depolarized potential (Colmers et al., 1982; see Materials and Methods, Whole-cell voltage-clamp and current-clamp recording from cultured bag cell neurons). H2O2 was tested at 30 μm, 100 μm, 300 μm, and 1 mm (n = 5, 8, 6, 7). The resulting inward current was concentration-dependent, with a threshold of 100 μm and the largest response at 1 mm (Fig. 2A,B). Fitting the concentration-response relationship with a four-parameter dose–response equation yielded an apparent EC50 of 14.2 mm with a Hill slope of 0.64. If this was due to an agonist-type mechanism, it would be expected that the H2O2-induced current would recover following drug removal. When the H2O2-containing nASW superfusing over the soma was replaced with nASW alone, shortly after the H2O2-induced current reached peak, it recovered by ∼75% (n = 4; Fig. 2C, inset).
A concentration-dependent H2O2-induced inward current. A, Current responses to bath-applied 30 μm, 100 μm, 300 μm, or 1 mm H2O2 (at bar) for different cultured bag cell neurons whole-cell voltage-clamped at −30 mV in nASW with Cs+-Asp-based intracellular saline. Scale bars apply to all traces. B, Concentration–response curve reveals increasingly higher amounts of H2O2 induces progressively larger currents (30 μm = 4.0 ± 7.0 pA, 100 μm = 66.0 ± 13.0 pA, 300 μm = 94.0 ± 27.0 pA, 1 mm H2O2 = 218.0 ± 66.0 pA). The line represents the fit of the data with a four-parameter dose–response equation, and provides an EC50 of 14.2 mm with a Hill slope of 0.64. C, Upon washout of 1 mm H2O2 with nASW, at the height of the response, the current recovered largely back to the baseline. Inset, Summary graph shows 73.0 ± 10.6% recovery, calculated by comparing the baseline (before H2O2 superfusion) and the stable current at end of the trace.
The H2O2-induced current is consistent with the opening of a nonselective cation channel
To verify channel opening, the membrane conductance was examined before and after H2O2 application, by delivering a 200 ms step to −40 mV from a holding potential of −30 mV under voltage-clamp (Fig. 3A, middle). A first step was delivered after a 15 min dialysis period, while a second step was given ∼90 s after that, right before 1 mm H2O2 superfusion, and a third step was performed at the peak of the H2O2 response (Fig. 3A, top). The ratios of the current from the second step (just before H2O2) versus the first step (Fig. 3B, control 2/1), and the third step (at the peak of H2O2 response) versus second step (Fig. 3B, H2O2 3/2) were taken to ascertain the change in conductance. The period preceding H2O2 presented essentially no change in conductance (∼0.99-fold); however, the conductance change at the peak of the H2O2 response was significantly larger (by ∼1.75-fold), in agreement with channel opening (n = 8; Fig. 3A, bottom, B).
H2O2 increases membrane conductance. A, Top, Typical example of 1 mm H2O2 superfusion (at bar) evoking a current in a bag cell neuron whole-cell voltage-clamped at −30 mV in nASW using Cs+-based intracellular saline. Middle, To determine membrane conductance, a 200 ms step to −40 mV is given ∼90 s before H2O2 (arrow, circled 1). A second step is taken right before H2O2 delivery (circled 2), whereas a third step is taken at the peak of the response (circled 3). Bottom, The current evoked by the step is markedly elevated at the peak of the H2O2 current (3; black trace) compared with that taken immediately (2; dark gray) or ∼90 s before H2O2 (1; light gray). B, Summary graph shows a significant increase in fold-change conductance, calculated by obtaining the ratios of the step-current taken just before H2O2 superfusion versus ∼90 s earlier (control 2/1 = 1.0 ± 0.079) and during H2O2 versus just before superfusion (H2O2 3/2 = 1.7 ± 0.33; r = 0.02381; *p = 0.0078, Wilcoxon matched-pairs signed ranks test).
We next characterized the ionic basis of the current. The permeability of Na+ and Ca2+ was studied by replacing extracellular Na+ with NMDG and/or Ca2+ with Mg2+ (see Materials and Methods, Whole-cell voltage-clamp). Compared with nASW as control (n = 11), the H2O2-induced current was reduced by ∼87% in Ca2+-free (n = 12), ∼95% in Na+-free (n = 6), and was almost absent in Na+/Ca2+-free (n = 8) extracellular saline (Fig. 4A), suggesting that H2O2 acts on a nonselective cation conductance. On average, the observed decreases in current were significant versus control (Fig. 4B).
The H2O2-induced current is sensitive to change in extracellular cations. A, Different individual cultured bag cell neurons under whole-cell voltage-clamp at −30 mV using Cs+-based intracellular saline. Compared with the current elicited by 1 mm H2O2 (at bar) in Na+-containing nASW, the response is smaller in Ca2+-free or Na+-free saline, and mostly eliminated in Na+/Ca2+-free saline. Scale bars apply to all traces. B, Group data show that, in contrast to nASW (1.2 ± 0.2 nA), Ca2+-free (151.6 ± 26.4 pA), Na+-free (65.3 ± 15.3 pA), or Na+/Ca2+-free saline (14.1 ± 7.6 pA) all significantly reduce the current (H = 29.991; df = 2; p < 0.0001, KW-ANOVA; *p < 0.05 nASW vs Ca+-free; *p < 0.01 nASW vs Na+-free; *p < 0.001, nASW vs Na+/Ca2+-free, Dunn's multiple-comparisons test). C, Difference currents obtained by subtracting the response to a 5 s, −60 to +60 mV voltage ramp (bottom inset), taken immediately before 1 mm H2O2 application, from that taken at the peak of the response. For nASW (black trace), the current is voltage-dependent and reverses between +30 and +40 mV. Upper insets show magnified Ca2+-free (dark gray), Na+-free (medium gray), and Na+/Ca2+-free (light gray) difference currents from −45 to −15 mV (left) and −15 to +15 mV (right). D, Average data illustrates a significant left-shift in reversal potential from nASW (30.6 ± 4.1 mV) with Ca2+-free (−8.4 ± 3.4 mV), Na+-free (−11.3 ± 3.4 mV), or Na+/Ca2+-free (−10.3 ± 3.7 mV) saline (F(3,37) = 27.329; p < 0.0001, ordinary ANOVA; *p < 0.01, nASW vs Ca2+-free, nASW vs Na+-free, and nASW vs Na+/Ca2+-free saline, Dunnett multiple-comparisons test). Bars as per panel B.
Ion substitution was also used to investigate the reversal potential of the H2O2-induced current. Specifically, a 5 s ramp from −60 to +60 mV was delivered under voltage-clamp from a holding potential of −30 mV (Fig. 4C). The ramp was given twice, i.e., right before H2O2 application, and again at the peak of the H2O2 response. A difference current was then calculated by subtracting the first ramp-induced current from the second ramp-induced current. In nASW, the difference current was nonlinear (inward between ∼−30 and ∼+30 mV), voltage-dependent, and reversed at ∼+30 mV (n = 10; Fig. 4C). However, in the absence of extracellular Ca2+, or Na+, or Na+/Ca2+, the difference currents flattened out (Fig. 4C, top insets). Moreover, the reversal potential was significantly left-shifted in Ca2+-free (∼−8 mV; n = 15), Na+-free (∼−11 mV; n = 6), or Na+/Ca2+-free saline (∼−10 mV; n = 10), as expected for a nonselective cation conductance (Fig. 4D).
The pharmacology of the H2O2-induced current is also consistent with a cation channel
Because the ion substitution and reversal potential results suggested a nonselective cation channel, we subsequently tested known cation channel blockers. In particular, 9-Pt, a purported transient receptor potential (TRP) cation channel melastatin subfamily isoform 4-specific inhibitor (Grand et al., 2008; Guinamard et al., 2014), clotrimazole, a general cation channel blocker known to prevent H2O2-induced currents in striatal and hippocampal neurons (Hill et al., 2004; Olah et al., 2009), and SKF-96365, which is often reported as a TRP channel canonical subfamily antagonist (Zhu et al., 1998), were used. Either 9-phenanthrol or clotrimazole were given at the peak of the response to 1 mm H2O2, the prediction being that a blocker would inhibit the channel and, compared with the vehicle (DMSO), increase the percent-recovery of the current (see Materials and Methods, Analysis). Whereas DMSO application (the vehicle; n = 4, 4) resulted in 10–30% recovery (Fig. 5A,C), introducing 100 μm 9-phenanthrol (n = 6) or 10 μm clotrimazole (n = 4; Fig. 5B, top, bottom), significantly increased the percent-recovery of the H2O2-induced current to ∼75 and ∼55%, respectively (Fig. 5C,D). For SKF-96365, 10 μm of the blocker was bath-applied 10 min before 1 mm H2O2 delivery; however, compared with H2O2 alone, the presence of SKF-96365 did not significantly alter the current (Fig. 5E).
The pharmacology of the H2O2-induced current is consistent with a cation channel. A, Whole-cell voltage-clamp recording from a cultured bag cell neuron in nASW at −30 mV using Cs+-Asp-based intracellular saline shows an inward current in response to 1 mm H2O2 (at bar). Bath-application of 0.1% (v/v) DMSO (vehicle, at second bar) at the peak of the response fails to alter the normal recovery of the current. B, Delivery of 100 μm 9-Pt (top) or 10 μm clotrimazole (bottom) at the peak of the response markedly inhibits the ongoing 1 mm H2O2-induced current. C, D, Summary graphs show both 9-Pt and clotrimazole significantly increase the percent recovery of the H2O2-induced current to 76.0 ± 11.4% and 56.5 ± 14.1%, respectively, calculated by comparing the baseline (before H2O2 bath-application) and the steady-state current at the end of the trace (t(8) = 2.743; *p = 0.0253, 9-Pt, t(6) = 3.183; *p = 0.0190, clotrimazole, both unpaired Student's t test). E, Group data of the H2O2-induced current in control vs neurons pretreated with 10 μm SKF-96365. In contrast to 1 mm H2O2 alone (93.4 ± 11.4 pA), the presence of 10 μm SKF-96365 (73.3 ± 11.2 pA) does not significantly reduce the response (U(5,12) = 21.0; p = 0.3827 H2O2 vs H2O2 in SKF-96365; Mann–Whitney U test).
Some TRPC channels are sensitive to changes in glucose concentration; for example, Lee et al. (2015b) reported that low glucose enhances neuronal TRPC channels, likely through a pathway involving the triggering of adenosine monophosphate-activated protein kinase. In the present study, the extracellular saline normally did not contain glucose (see Materials and Methods, Whole-cell voltage-clamp and current-clamp recording from cultured bag cell neurons); thus, we examined whether introducing 1 mm extracellular glucose beforehand would impact the H2O2-induced current. The presence of this low glucose concentration did not significantly alter the response to 1 mm H2O2 (H2O2 alone: 65.3 ± 16.0 pA, n = 4 vs H2O2 in glucose: 68.0 ± 16.7 pA, n = 4; t(6) = 0.1189; p = 0.9093, unpaired Student's t test). Because the Cs+-based intracellular saline used for most whole-cell experiments contained 11 mm glucose, we also performed a control for this intracellular addition of glucose via the pipette by recording the H2O2-induced current using sharp-electrode voltage-clamp. Neurons were impaled with a sharp-electrode (which does not permit appreciable exchange with the cytosol) and held at −30 mV (see Materials and Methods, Sharp-electrode recording from bag cell neurons in culture and isolated clusters). We found that the current under the sharp-electrode configuration (140 ± 39.2 pA; n = 6) was not significantly different from parallel controls performed using whole-cell pipettes containing Cs+-based intracellular saline with glucose (112 ± 32.2 pA; n = 5; t(9) = 0.5390; p = 0.6030, unpaired Student's t test).
Our data suggest that the H2O2-induced current is mediated by a conductance similar to the voltage-dependent nonselective cation channel characterized in bag cell neurons by Wilson et al. (1996). Those authors reported that this channel was blocked by relatively high concentrations (80–100 μm) of the classical Na+ channel antagonist, TTX (Narahashi et al., 1964). Thus, to confirm any similarity, we initially bath-applied 100 μm TTX at the peak of H2O2-induced current. However, we observed no difference in percent-recovery of the response in the presence (n = 5; 41.6 ± 7.7%) or absence of TTX (n = 6; 41.2 ± 8.6%; t(9) = 0.03097; p = 0.9760, unpaired Student's t test). TTX has one primary alcohol and three secondary alcohol functional groups. Chicheportiche et al. (1980) showed that the secondary alcohols can be oxidized, resulting in inactive toxin. Thus, in the present study TTX may have been oxidized upon bath-application in the presence of H2O2. To avoid this, we opted for pretreatment with TTX and delivering H2O2 afterward. Compared with the control H2O2-induced current (n = 10), a 30 min pretreatment with 100 μm TTX (n = 7) reduced the response by ∼40% (Fig. 6A, top, middle). Moreover, a further decrease in the response, to ∼15% of control, was seen with 300 μm TTX pretreatment (n = 4; Fig. 6A, bottom). The presence of 100 or 300 μm TTX significantly reduced the H2O2 response versus control (Fig. 6B).
The H2O2-induced current is reduced by tetrodotoxin. A, Current responses to 1 mm H2O2 (at bar) of separate cultured bag cell neurons whole-cell voltage-clamped at −30 mV in nASW with Cs+-based intracellular saline. Compared with control (top), a 30 min pretreatment with 100 μm TTX (middle) noticeably reduces the H2O2-induced current, while 300 μm, TTX (bottom) almost eliminates the response. Ordinate applies to all traces. B, Group data of the H2O2-induced current in 100 or 300 μm TTX vs control. Compared with 1 mm H2O2 alone (79.9 ± 7.7 pA), the presence of 100 μm (47.1 ± 9.4 pA) or 300 μm TTX (15.4 ± 3.6 pA) significantly reduces the response (F(2,18) = 12.570; p = 0.0004 ordinary ANOVA, *p < 0.05 H2O2 vs H2O2 post 100 μm TTX; *p < 0.01 H2O2 vs H2O2 post 300 μm TTX, Dunnett multiple-comparisons test).
The H2O2-induced current is boosted by preventing oxidation and attenuated by promoting reduction
Glutathione peroxidase is a cytosolic enzyme that catalyzes the reduction of H2O2 to H2O, and the concomitant oxidation of glutathione (Jones et al., 1981). Mercaptosuccinate inhibits glutathione peroxidase, thus preventing H2O2 reduction (Dringen et al., 1998); hence, we hypothesized that delivering mercaptosuccinate for ∼10 min before H2O2 would increase the response. Bath-application of 1 mm mercaptosuccinate alone had little to no effect on the holding current (n = 10; Fig. 7A, top). As expected, introducing 1 mm H2O2 on its own (n = 6) produced an inward current; in addition, this was significantly enhanced by ∼50% in the presence of mercaptosuccinate (n = 8; Fig. 7A, middle, bottom, B). To further examine the role of glutathione in moderating the H2O2-induced current, we initially exposed neurons to 100 μm of N-acetylcysteine, which increases glutathione levels by elevating the intracellular concentration of the glutathione synthesis precursor, cysteine (Cotgreave et al., 1991). On its own, N-acetylcysteine did not cause a change in current at −30 mV (n = 6; Fig. 7C, top); yet, delivery of 1 mm H2O2 ∼10 min after N-acetylcysteine (n = 6) evoked an inward current which was significantly reduced by ∼40% when contrasted with the response to just H2O2 (n = 7; Fig. 7C, middle, bottom, D).
The H2O2-induced current is enhanced by mercaptosuccinate and reduced by N-acetylcysteine or dithiothreitol. Cultured bag cell neurons are bathed in nASW and whole-cell voltage-clamped at −30 mV using Cs+-based intracellular saline. A, Bath-application (at bar) of 1 mm mercaptosuccinate alone has little effect (top), whereas delivery of 1 mm H2O2 to a second cell elicits a noticeable inward current (middle). Moreover, in a third neuron initially given mercaptosuccinate (mercapto), the H2O2-induced current is enhanced by ∼50% (bottom). Ordinate applies to all traces. B, Summary graph of peak mercaptosuccinate- and H2O2-induced current ± mercaptosuccinate. There is a significant difference between both the H2O2-induced current with (98.1 ± 9.6 pA) and without (66.4 ± 5.8 pA) mercaptosuccinate as well as mercaptosuccinate alone (7.3 ± 7.2 pA; F(2,21) = 50.879; p < 0.0001, ordinary ANOVA; *p < 0.001 mercapto vs H2O2; *p < 0.05 H2O2 vs H2O2 in mercapto, Tukey–Kramer multiple-comparisons test). C, Introducing 100 μm NAC has only a nominal impact (top), whereas 1 mm H2O2 generates an inward current (middle). In the presence of N-acetylcysteine, the H2O2-induced current is reduced by ∼40% (bottom). D, Summary data showing that compared with N-acetylcysteine alone (43.3 ± 25.7 pA), the H2O2-induced current differs significantly with (90.3 ± 21.4 pA) and without N-acetylcysteine (148.0 ± 16.0 pA; U(6,7) = 0; *p = 0.0034 NAC vs H2O2; U(6,7) = 5; *p = 0.0221 H2O2 vs H2O2 in NAC, both Mann–Whitney U test). E, Minimal response to 1 mm DTT alone (top), whereas 1 mm H2O2 again elicits a clear inward current (middle); in addition, when DTT is already present, the H2O2-induced current is reduced by ∼70% (bottom). F, Group data showing that compared with DTT alone (11.3 ± 4.5 pA), the H2O2-induced current differs significantly with (43.2 ± 9.6 pA) and without DTT (155.0 ± 13.2 pA; F(2,15) = 73.831; p < 0.0001, ordinary ANOVA; *p < 0.001 DTT vs H2O2; *p < 0.001 H2O2 vs H2O2 in DTT, Tukey–Kramer multiple-comparisons test).
To test whether cation channel activation by H2O2 occurs via oxidation of sulfhydryl groups, we used DTT, a reducing agent that maintains thiol groups on amino acids, such as Cys, in the reduced state (Cleland, 1964). Application of DTT alone did not elicit a current at −30 mV (n = 6; Fig. 7E, top). However, in the presence of DTT, the response to 1 mm H2O2 was obviously decreased compared with the delivery of H2O2 alone (n = 6; Fig. 7E, middle, bottom). The H2O2 current was reduced by ∼70% and this drop was significant (Fig. 7F).
NEM- and H2O2-induced currents are occlusive
The ability of DTT to prevent H2O2 from activating the current suggests that H2O2 may gate the cation channel by direct oxidation of sulfur-containing amino acids, in particular, Cys and/or Met residues in the channel or an associated protein(s) (Hoshi and Heinemann, 2001). NEM is a sulfydryl alkylating agent (Jakobs et al., 1982) that can remove hydrogen from thiol side chains and replace it with a carbon–sulfur bond, potentially minimizing H2O2-mediated oxidation (Gregory, 1955). Therefore, we tested the effects of NEM on both membrane current itself and the H2O2-induced current. Bath-application of 300 μm NEM elicited an ∼200 pA inward current at −30 mV (n = 9), with the subsequent addition of 1 mm H2O2 evoking only a small response of ∼12 pA (n = 9; Fig. 8A). Similarly, when H2O2 was added to the bath first, it brought about a typical inward current (n = 5), whereas delivery of NEM afterward again caused little change (n = 5; Fig. 8B). The currents induced by H2O2 and NEM, respectively, were not significantly different on average (Fig. 8C,D, two white bars); moreover, the two currents occluded one another (Fig. 8C,D, two black bars).
NEM- and H2O2-induced currents are occlusive. Representative response to 300 μm NEM or 1 mm H2O2 of different cultured bag cell neurons whole-cell voltage-clamped at −30 mV in nASW with Cs+-based intracellular saline. A, Bath-application (at bar) of NEM evokes a prominent inward current. Yet, H2O2 (at second bar) in the presence of NEM, yields little further change. B, Delivery of H2O2 elicits a typical inward current in a separate neuron, but there is no obvious current when NEM is applied in the presence of H2O2. C, D, Summary data demonstrates a significant difference between the NEM-elicited current (201.0 ± 65.5 pA) and the H2O2-induced current post-NEM (11.7 ± 13.1 pA; U(9,9) = 7.0; *p = 0.0019, Mann–Whitney U test). Similarly, there is a significant difference between the H2O2-induced current (129.6 ± 24.5 pA) and the NEM-elicited current post-H2O2 (6.3 ± 5.9 pA; (t(4) = 4.90; *p = 0.008, unpaired Student's t test, Welch corrected). Note, there is no statistical difference between the initial current produced by NEM and H2O2 (white bars; t(9) = 1.022; p = 0.3337, unpaired Student's t test, Welch corrected); as well, the NEM- and the H2O2-induced currents in the presence of H2O2 and NEM, respectively, are not statistically different (black bars; U(5,9) = 14.0; p = 0.2977, Mann–Whitney U test).
H2O2 depolarizes and induces robust action potential firing
Given that H2O2 application elicits an inward current, we sought to determine the impact of introducing H2O2 on cultured bag cell neuron membrane potential. This was performed using whole-cell current-clamp in nASW with the K+-based intracellular saline in the pipette. H2O2 was tested at both −40 mV, the approximate membrane potential during the afterdischarge, and −60 mV, the approximate membrane potential before or at the start of the afterdischarge (Kupfermann and Kandel, 1970; Kaczmarek et al., 1982). Bath application of 100 μm H2O2 induced a depolarization of ∼7 and ∼14 mV from −40 (n = 7) and −60 mV (n = 5), respectively, versus H2O as control at −40 mV (n = 5) or −60 mV (n = 5), which had no obvious effect (Fig. 9A,B). In no case did 100 μm H2O2 cause the neuron to fire action potentials. However, 1 mm H2O2 not only depolarized the membrane, by ∼11 mV for −40 mV (n = 5) and ∼22 mV for −60 mV (n = 5), but consistently evoked a burst of spikes (Fig. 9C). Compared with H2O, the magnitudes of the responses were significant for both −40 and −60 mV (Fig. 9D). As for the action potential firing, it lasted for ∼8 min at −40 mV and ∼7 min at −60 mV with the latency of ∼3 min and frequency of ∼1 Hz at both voltages (Fig. 9E).
H2O2 depolarizes and induces robust action potential firing. A, Separate cultured bag cell neurons are whole-cell current-clamped in nASW with K+-based intracellular saline and initially set to either −40 or −60 mV with bias current. Bath-application (at bar) of H2O (0.5% v/v) at −40 (top) or −60 mV (bottom) does not affect membrane voltage. Ordinate applies to both traces. B, Exposure to 100 μm H2O2 depolarizes two different neurons from both −40 (top) and −60 (bottom) mV, but neither reaches threshold. Ordinate applies to both traces. C, Delivery of 1 mm H2O2 depolarizes the membrane and induces robust action potential firing from both −40 (top) and −60 (bottom) mV in separate cells. Ordinate applies to both traces. D, Summary graph indicating the average depolarizations from −40 (open circles) or −60 (closed circles) mV. Compared with the response produced by applying H2O (−40 mV: 2.8 ± 0.8 mV; −60 mV: 1.1 ± 0.6 mV), the depolarization elicited by 100 μm H2O2 (−40 mV: 7.5 ± 0.6 mV; −60 mV: 14.0 ± 2.6 mV) or 1 mm H2O2 (−40 mV: 11.4 ± 2.9; −60 mV: 21.8 ± 2.1) is significantly different (−40 mV: H = 8.545; df = 1; p = 0.0070, KW-ANOVA; p > 0.05 H2O vs 100 μm H2O2; *p < 0.05 H2O vs 1 mm H2O2, Dunn's multiple-comparison's test; −60 mV: F(2,12) = 29.275; p < 0.0001, ordinary ANOVA; *p < 0.01 H2O vs 100 μm H2O2, *p < 0.001 H2O vs 1 mm H2O2, Tukey–Kramer multiple-comparisons test). E, Group data show no significant difference in the frequency (t(8) = 0.2841; p = 0.7835, unpaired Student's t test) or duration (t(8) = 0.3884; p = 0.7079, unpaired Student's t test) of action potential firing from −40 versus −60 mV caused by 1 mm H2O2. In addition, there is no significant difference in the latency, i.e., the time it takes for the neuron to start firing action potentials post-H2O2 delivery, between responses at −40 mV (2.8 ± 0.6 mV) and −60 mV (3.0 ± 0.6 mV; t(8) = 0.3001; p = 0.7718, unpaired Student's t test).
Knowing that the H2O2-induced current is blocked by 9-phenanthrol, clotrimazole, or TTX, we next investigated whether these blockers altered the depolarization to 1 mm H2O2 from −40 mV. As expected, 1 mm H2O2 (n = 18) elicited a depolarization of ∼9 mV and action potential firing of ∼1 Hz that lasted for ∼12 min (Fig. 10A, top, B). This response was essentially eliminated by a 20 min pretreatment with either 10 μm clotrimazole (n = 6; −2.8 ± 1.4 mV) or 100 μm 9-phenanthrol (n = 6; 4.0 ± 2.2 mV; Fig. 10A, middle). As for TTX (n = 4), incubating in 300 μm did not prevent the depolarization altogether (Fig. 10B, top); instead, it significantly reduced the firing frequency to ∼0.5 Hz and the duration of spiking to ∼3 min (Fig. 10A, bottom, B, middle, bottom).
H2O2-evoked spiking is reduced by tetrodotoxin and prevented by 9-Pt or clotrimazole. Voltage responses to 1 mm H2O2 of different cultured bag cell neurons whole-cell current-clamped to −40 mV in nASW with K+-based intracellular saline after 20 min pretreatment with clotrimazole, 9-Pt, or TTX. A, Bath-application of 1 mm H2O2 (at bar) induces robust action potential firing (top). However, a 20 min pretreatment with 10 μm clotrimazole (upper middle) or 100 μm 9-phenanthrol (lower middle) virtually eliminates the depolarization and prevents firing all together. Finally, after 20 min of 300 μm TTX (bottom), introducing H2O2 still leads to membrane depolarization, but the duration and frequency of action potential firing is lessened. The ordinate applies to all traces. B, Summary graph indicating the average H2O2-evoked depolarization (top) in the presence of 300 μm TTX does not differ significantly from 1 mm H2O2 alone (U(4,18) = 32.0; p = 0.7743, Mann–Whitney U test). However, spike frequency (middle) and duration (bottom) are significantly reduced when TTX is in the bath (frequency: U(4,18) = 6.0; *p = 0.0120; duration: U(4,18) = 5.0; *p = 0.0049, both Mann–Whitney U test).
H2O2 depolarizes bag cell neurons and initiates an afterdischarge in desheathed clusters
Because H2O2 both gates a cation channel and causes bursting in cultured bag cell neurons, we sought to ascertain whether H2O2 can produce afterdischarge-like responses from the bag cell neuron cluster itself. To begin with, 1–10 mm H2O2 was bath-applied to the entire abdominal ganglion while recording extracellularly from one of the two intact bag cell neuron clusters (see Materials and Methods, Ensemble extracellular recording from the intact bag cell neuron cluster). However, this failed to bring about an afterdischarge nor did H2O2 affect the ability of synaptic input to evoke an afterdischarge, as subsequent stimulation of the pleuroabdominal connective resulted in normal afterdischarges that were of the same duration (n = 4; 57.8 ± 12.6 min) compared with those elicited in ganglia not exposed to H2O2 (n = 7; 44.4 ± 14.1 min; t(9) = 0.6349; p = 0.5413, unpaired Student's t test).
It is possible that the sheath surrounding the cluster hindered diffusion of H2O2 to the bag cell neurons and/or antioxidant mechanisms present either in the sheath cells or the extracellular space reduced the exogenous H2O2. Therefore, to demonstrate that direct delivery of H2O2 to the cluster can elicit bursting, the connective sheath was removed and the entire bag cell neuron cluster was isolated from the abdominal ganglion and placed in a culture dish. Sharp-electrode recordings were then made within the cluster from individual bag cell neurons initially current-clamped at −60 mV (see Materials and Methods, Sharp-electrode recording from bag cell neurons in culture and isolated clusters). Still, bath-applying 10 mm H2O2 alone resulted in a depolarization of just ∼10 mV (n = 5; Fig. 11C, left, white bar), with only one of the five preparations exhibiting a 10.9 min burst of action potentials.
H2O2 depolarizes bag cell neurons and initiates an afterdischarge in desheathed clusters. A, Under sharp-electrode current-clamp, a 2 s 1 mm ACh pressure ejection (arrow) to one side of the cluster depolarizes a bag cell neuron recorded on the opposite side. B, Bath-application of 10 mm H2O2 (bar) post-ACh delivery initiates a prolonged burst in a bag cell neuron within a cluster. Inset, Phase-contrast photomicrograph of a desheathed bag cell neuron cluster in tcASW indicating the placement of the ACh-containing pressure ejection electrode and the intracellular current-clamp (cc) sharp electrode. C, Following 1 mm ACh pressure ejection, 10 mm H2O2 is introduced 8.3 min later and induces a depolarization (23.2 ± 2.5 mV; t(12) = 3.375; *p = 0.0055, unpaired Student's t test) that evokes a burst, whereas 10 mm H2O2 alone solely depolarizes the neuron, but to a significantly lesser extent (9.8 ± 3.6 mV; left). During the H2O2-evoked burst, the frequency of action potential firing is 0.8 ± 0.2 Hz (left), mean discharge duration is 6.4 ± 1.4 min (middle), and the latency is 3.4 ± 0.7 min (right).
We speculated that H2O2 may need to work in concert with the known bag cell neuron input transmitter, ACh (White and Magoski, 2012). Previously, our lab showed that brief pressure-application of ACh to the isolated bag cell neuron cluster would initiate a transient depolarization in a given neuron, sometimes accompanied by afterdischarge-like spiking (White et al., 2018). Thus, in the present study, while again recording from a single bag cell neuron within the cluster, ACh was first pressure-ejected followed by the introduction of H2O2 (Fig. 11B, inset). In nine separate preparations, a 2 s pressure-ejection of 1 mm ACh to one side of the cluster caused depolarization of a bag cell neuron recorded on the other side, with a mean of 6.4 ± 3.4 mV (Fig. 11A). Moreover, bath-applying 10 mm H2O2, 8.3 ± 1.3 min later, evoked a burst within ∼3 min, consisting of a depolarization of ∼25 mV and spiking at ∼0.8 Hz for ∼6 min (Fig. 11B,C, black bars). In all nine clusters, administering H2O2 after ACh always led to spiking, as well as a significantly greater depolarization compared with the five clusters exposed to H2O2 alone (Fig. 11C).
Discussion
H2O2 can oxidize DNA, lipids, or proteins (Halliwell, 1992; Gutteridge and Halliwell, 1992), which may bring about synaptic plasticity (Kamsler and Segal, 2004; Kishida and Klann, 2007) or gate ion channels (Avshalumov et al., 2007). For example, various cation currents in hippocampal, nigral, and striatal neurons are opened by H2O2 (Hill et al., 2006; Olah et al., 2009; Lee et al., 2013). The bag cell neuron afterdischarge is maintained by various nonselective cation channels. In the fast-phase, Ca2+ influx and release opens a voltage-independent cation channel, with a linear current/voltage relationship and an ∼−40 mV reversal potential (Hung and Magoski, 2007; Hickey et al., 2010). With the slow-phase, phospholipase C cleaves PIP2 into IP3 and DAG (Fink et al., 1988). DAG gates a second, voltage-independent cation channel, with a distinct pharmacology and reversal potential of ∼−20 mV (Sturgeon and Magoski, 2016). This channel may be similar to TRPC3/6/7 channels, which are well established as being activated by DAG (Hofmann et al., 1999; Okada et al., 1999). DAG also activates PKC to mediate a number of afterdischarge-associated events (DeRiemer et al., 1984; Wayne et al., 1999; Groten and Magoski, 2015), including the regulation of a third cation channel that is voltage-dependent with a reversal potential >+30 mV (Wilson et al., 1996, 1998; Magoski and Kaczmarek, 2005; Gardam and Magoski, 2009; Sturgeon and Magoski, 2018). PKC also stimulates H2O2 production in bag cell neurons (Munnamalai et al., 2014), suggesting that H2O2 may be a signaling molecule during the afterdischarge.
We find that extracellular application of H2O2 elicits a prolonged voltage- and concentration-dependent inward current in cultured bag cell neurons. The response is observed with as low as a 100 μm H2O2, although for consistency we primarily used 1 mm. Previous reports involving both cultured Aplysia sensory neurons and various mammalian neurons in vitro or ex vivo use comparable amounts of H2O2 and, similar to our work, show channel gating but no obvious damage (Chen et al., 2001; Chang et al., 2003; Olah et al., 2009; Lee et al., 2011, 2013; Ohashi et al., 2016). Quite possibly, these concentrations are necessary because the fast consumption of H2O2 by intracellular antioxidant enzymes produces a high-extracellular to low-intracellular H2O2 gradient, estimated by some to be up to 650-fold (Huang and Sikes, 2014). Thus, it is plausible that H2O2 falls substantially as it is scavenged once diffusing across the bag cell neuron membrane.
The increase in membrane conductance during the H2O2-induced current is consistent with channel opening, whereas the ∼+30 mV reversal potential in Na+-containing nASW suggests the conductance is nonselective for cations. Depending on specific ionic permeability, cation channels can have a wide range of reversal potentials (−40 to +50 mV; Partridge et al., 1994; Clapham, 2003). The bag cell neuron H2O2-induced current is dependent on select extracellular cations, i.e., the current is lessened in the absence of extracellular Ca2+, diminished even further whether only Na+ is removed, and nearly eliminated when both cations are taken away. In addition, the Ca2+-, Na+-, or Na+/Ca2+-free external salines all negatively shift the reversal potential to ∼−10 mV, and drastically reduce any outward current at more positive voltages. Likely, the channel normally passes more Na+ and Ca2+ into the neuron than it does K+ out.
The U-shaped current/voltage relationship and ∼+30 mV reversal potential of the H2O2-induced current are very similar to that of the bag cell neuron cation channel reported by Wilson et al. (1996). That channel is both sensitive to the removal of extracellular Ca2+, which lowers the conductance and left-shifts the reversal potential, and is modestly blocked by intracellular Mg2+ (Geiger et al., 2009). For the H2O2-gated channel, Na+- and/or Ca2+-free saline may result in an insufficiency of extracellular cations in the pore; if those cations normally repulse intracellular Mg2+, their absence could enhance the Mg2+ block and impede outward current. There are accounts of TRPC4/5 and TRPM2 channels passing limited outward current following replacement of extracellular Na+ and Ca2+ with NMDG (Schaefer et al., 2000, 2002; Kühn and Lückhoff, 2004; Blair et al., 2009).
The pharmacology of the H2O2-induced current also indicates a cation channel. Both the inward current and depolarization brought about by H2O2 are inhibited by 9-Pt, an ostensible TRPM4-specific inhibitor (Grand et al., 2008; Guinamard et al., 2014), which also blocks the voltage-dependent cation channel (Sturgeon and Magoski, 2013), as well as clotrimazole, a general cation channel blocker that has been found to inhibit TRPM2 (Hill et al., 2004). Furthermore, pretreatment with TTX, which Wilson et al. (1996) showed blocks the voltage-dependent cation channel in bag cell neurons, prevents the H2O2-induced current. TTX also impacts the action potential burst elicited by H2O2, which is again reminiscent of Wilson et al. (1996), who found that TTX attenuates the depolarization brought about by opening the voltage-dependent cation channel. It is unlikely that TTX is effecting Na+ channels, given that both our laboratory and others rarely observe voltage-gated Na+ current in cultured bag cell neurons (Nick et al., 1996; Magoski et al., 2000). Based on similar biophysical and pharmacological properties, the H2O2-induced current and the current characterized by Wilson et al. (1996) appear to be the same or at the very least share one or more channel subunits.
Lee et al. (2011) demonstrated that inhibiting glutathione peroxidase with mercaptosuccinate increases the intracellular H2O2 concentration of substantia nigra neurons. Our finding that the H2O2-induced current is enhanced by mercaptosuccinate suggests glutathione peroxidase normally catalyzes the reduction of H2O2 in bag cell neurons. Fittingly, the H2O2-induced current is attenuated by N-acetylcysteine, a precursor for glutathione synthesis (Cotgreave et al., 1991). These findings are in agreement with gating by redox, where greater glutathione bioavailability lessens the current by eliminating more H2O2. Last, the response is decreased with dithiothreitol, which likely opposes the effects of H2O2 by keeping sulfhydryl groups on key Cys and/or Met residues in the reduced state (Cleland, 1964). It is presently unknown whether redox reactions underlying activation are on associated protein(s) or directly on the channel, as is the case for inwardly rectifying K+ channels and various TRPC or TRP vanilloid channels (Bannister et al., 1999; Yoshida et al., 2006).
TRPM2 is expressed in neurons and microglia, and can be triggered by H2O2 (Perraud et al., 2001; Sano et al., 2001; Kraft et al., 2004; Tong et al., 2006). An increase in endogenous H2O2 with mercaptosuccinate elevates the firing rate of substantia nigra GABAergic neurons by activating a TRPM2-like channel (Lee et al., 2011, 2013), possibly by oxidation of sulfhydryl groups on TRPM2 Cys residues (Ogawa et al., 2016). Our findings suggest a similar effect in bag cell neurons. NEM, an alkylating agent that is reactive toward thiols (Jakobs et al., 1982), evokes a current which appears analogous to the H2O2-induced current. Furthermore, addition of H2O2 after NEM does not induce a response and vice versa. If H2O2 modifies sulfhydryl groups on Cys residues, it would be unable to induce a response post-NEM application. Based on the redox-type activation mechanism we propose, the Aplysia current may involve a TRPM2-like subunit. That stated, the block by 9-Pt also points to a role for TRPM4-like channels, the mammalian versions of which are, like the bag cell neuron H2O2-induced current, voltage-dependent (Nilius et al., 2003).
Delivery of H2O2 to the desheathed, isolated bag cell neuron cluster can elicit spiking. Because bath-applying H2O2 alone causes a burst in a minority of clusters, we speculate H2O2 may need to work in concert with ACh, a known input transmitter that gates an ionotropic receptor on bag cell neurons (White and Magoski, 2012; White et al., 2014). Pressure-ejecting ACh onto one side of the cluster depolarizes a neuron recorded on the opposite side; this is because of the transfer of cholinergic current through electrotonic coupling between neurons within the cluster (Kupfermann and Kandel, 1970; Dargaei et al., 2014; White et al., 2018). When H2O2 is bath-applied subsequent to the ACh pressure-ejection, it consistently provokes an afterdischarge-like burst. ACh may facilitate H2O2-dependent opening of the voltage-dependent cation channel, either through direct depolarization or initiating an as yet unidentified intracellular pathway. Work by our laboratory and others indicates that if one bag cell neuron is bursting within the cluster, all other neurons also spike synchronously (Kupfermann and Kandel, 1970; Brown and Mayeri, 1989; Dargaei et al., 2014). Thus, it is reasonable to assume that the H2O2-induced afterdischarge-like response represents en masse firing of the cluster.
Activating PKC in bag cell neurons is sufficient to turn on H2O2 production (Munnamalai et al., 2014). Given that PKC is triggered during the afterdischarge (Wayne et al., 1999), and H2O2 exposure depolarizes bag cell neurons, as well as elicits prolonged action potential firing similar to the afterdischarge, it is possible that H2O2 is a signaling molecule for the afterdischarge. The H2O2-induced nonselective cation current may maintain the afterdischarge, thereby ensuring egg-laying hormone secretion and reproductive behavior. Thus, in addition to the link between oxidative damage and various neurodegenerative diseases (Brieger et al., 2012), the production of reactive oxygen species can have a role in physiological processes.
Footnotes
This work was supported by a Canadian Institutes of Health Research operating Grant (MOP111211) and project Grant (PJT-159794) to N.S.M. We thank C. A. London for technical support, and C. B. Beekharry and C. J. Groten for preliminary assistance with the H2O2-induced response.
The authors declare no competing financial interests.
- Correspondence should be addressed to Neil S. Magoski at magoski{at}queensu.ca
