Nitrergic modulation of neuronal excitability in the mouse hippocampus is mediated via regulation of Kv2 and voltage‐gated sodium channels

Regulation of neuronal activity is a necessity for communication and information transmission. Many regulatory processes which have been studied provide a complex picture of how neurons can respond to permanently changing functional requirements. One such activity‐dependent mechanism involves signaling mediated by nitric oxide (NO). Within the brain, NO is generated in response to neuronal NO synthase (nNOS) activation but NO‐dependent pathways regulating neuronal excitability in the hippocampus remain to be fully elucidated. This study was set out to systematically assess the effects of NO on ion channel activities and intrinsic excitabilities of pyramidal neurons within the CA1 region of the mouse hippocampus. We characterized whole‐cell potassium and sodium currents, both involved in action potential (AP) shaping and propagation and determined NO‐mediated changes in excitabilities and AP waveforms. Our data describe a novel signaling by which NO, in a cGMP‐independent manner, suppresses voltage‐gated Kv2 potassium and voltage‐gated sodium channel activities, thereby widening AP waveforms and reducing depolarization‐induced AP firing rates. Our data show that glutathione, which possesses denitrosylating activity, is sufficient to prevent the observed nitrergic effects on potassium and sodium channels, whereas inhibition of cGMP signaling is also sufficient to abolish NO modulation of sodium currents. We propose that NO suppresses both ion channel activities via redox signaling and that an additional cGMP‐mediated component is required to exert effects on sodium currents. Both mechanisms result in a dampened excitability and firing ability providing new data on nitrergic activities in the context of activity‐dependent regulation of neuronal function following nNOS activation.

oxide (NO). Within the brain, NO is generated in response to neuronal NO synthase (nNOS) activation but NO-dependent pathways regulating neuronal excitability in the hippocampus remain to be fully elucidated. This study was set out to systematically assess the effects of NO on ion channel activities and intrinsic excitabilities of pyramidal neurons within the CA1 region of the mouse hippocampus. We characterized whole-cell potassium and sodium currents, both involved in action potential (AP) shaping and propagation and determined NO-mediated changes in excitabilities and AP waveforms. Our data describe a novel signaling by which NO, in a cGMPindependent manner, suppresses voltage-gated Kv2 potassium and voltage-gated sodium channel activities, thereby widening AP waveforms and reducing depolarization-induced AP firing rates. Our data show that glutathione, which possesses denitrosylating activity, is sufficient to prevent the observed nitrergic effects on potassium and sodium channels, whereas inhibition of cGMP signaling is also sufficient to abolish NO modulation of sodium currents. We propose that NO suppresses both ion channel activities via redox signaling and that an additional cGMP-mediated component is required to exert effects on sodium currents. Both mechanisms result in a dampened excitability and firing ability providing new data on nitrergic activities in the context of activity-dependent regulation of neuronal function following nNOS activation.
K E Y W O R D S excitability, hippocampus, nitric oxide, redox signaling, voltage-gated potassium channel, voltage-gated sodium channel 1 | INTRODUCTION Regulation of neuronal intrinsic excitability is an essential mechanism in adapting neuronal function to continuously changing requirements for reliable information transmission. Many activity-dependent mechanisms have been reported in various neuronal cell types and networks where homeostatic control of network activity can occur via multiple signaling pathways. One example includes changes in synaptic strength, a phenomenon termed "synaptic scaling" (Turrigiano, 1999) with underlying key mechanisms include changes in (i) synaptic activity-driven calcium signaling via glutamate receptor activation, (ii) voltage-gated calcium entry, or (iii) potassium conductances (Brickley et al., 2001;O'Leary et al., 2010). Following calcium influx, an important cascade is triggered, namely the activation of calcium/calmodulin-dependent neuronal nitric oxide (NO) synthase (nNOS). Resulting NO production activates broad signaling, including the NO-induced activation of the canonical cGMP pathway mediated by cGMP generation, activation of protein kinase G (PKG) and associated downstream phosphorylation events. In addition, NO induces the direct posttranslational protein modifications S-nitrosylation (a nonenzymatic reversible reaction of NO À with cysteine residues) and 3-nitrotyrosination (an irreversible reaction of a tyrosine residue with NO 2 ). These effects have been widely reported under physiological and neuropathological conditions (Akhtar et al., 2012;Lipton et al., 1994). Both redox-sensitive modifications impact on protein function and thus are essential regulators of neuronal activity in health and disease (Balez & Ooi, 2016;Bradley & Steinert, 2016;Garthwaite, 2019;. Modulation by NO has been reported for various ion channels, including voltagegated potassium channels (Kv), the M-current (Kv7), calcium, sodium, and TRP channels (Gamper & Ooi, 2015). Suppression of Kv3 currents in CHO cells is mediated via cGMP production and phosphorylation (Moreno et al., 2001) and cGMP-mediated suppression of native Kv3.1 currents was shown in principal neurons of the medial nucleus of the trapezoid body (MNTB) (Steinert et al., 2008). L-type calcium channels are inhibited by NO in a cGMP-dependent manner in myofibroblasts (Bae et al., 2020) but augmented in principal neurons of MNTB (Tozer et al., 2012).

It has been long established that nNOS is expressed in hippocampal
GABAergic nonprincipal neurons (Jinno et al., 1999;Megias et al., 1997) with a relatively large subpopulation (approximately 30%) of hippocampal parvalbumin-and neuropeptide Y-positive GABAergic interneurons being nNOS positive (Jinno et al., 1999;Tricoire et al., 2010). In addition to regional difference in nNOS expression, functional NO release across hippocampal regions differs greatly (Lourenco, Ferreira, et al., 2014) indicating the diversity of NO production and bioactivity.
Despite published work on activity-dependent homeostasis and modulation of intrinsic excitability, a comprehensive understanding of the mechanisms initiated by NO on neuronal properties is currently lacking due to the diversity of experimental preparations and applications of NO (Àdonors).
Given the importance and complexity of the nitrergic signaling route, this study was set out to characterize the effects of NO on ion channel activities and excitabilities in native pyramidal neurons within the CA1 hippocampal region by a defined concentration of NO to allow a systematic assessment of nitrergic effects. CA1 pyramidal neurons possess a range of Kv channels. The main determinants in shaping the action potential (AP) waveform and regulating excitabilities are potassium channels belonging to the families of high-voltage gated Kv3 (mainly Kv3.1 and Kv3.2) and Kv2 "delayed-rectifier" channels, with Kv2.1 being the predominant Kv channel expressed in pyramidal neurons (Kirizs et al., 2014). Low voltage-activated Kv1 channels are involved in setting AP thresholds (Jan & Jan, 2012) and voltage-gated sodium channels (Na v ), once activated, are responsible for the depolarizing upstroke and amplitude of an AP (Bean, 2007). The most prominent members of the Na v family in hippocampal pyramidal neurons are the Na v 1.2 and Na v 1.6 (Westenbroek et al., 1989). In addition, there are several potassium channels possessing a more regulatory function which are activated around resting membrane potentials, such as the hyperpolarization-activated cyclic nucleotide-gated ion channel (HCN) or the M-current (belonging to the subthreshold-active Kv7 family of potassium channels). For instance, Kv7.2/Kv7.3 channels stabilize the membrane potential and increase Na v channel availability and AP amplitudes (see review Trimmer, 2015).
In this study, we assessed the main elements of information transmission in native CA1 pyramidal neurons, for example, AP firing and neuronal excitability, which are predominantly regulated by Kv2 and voltage-gated sodium channels. We characterized the involvement of cGMP-and redox-sensitive signaling by applying voltage protocols which specifically assess high voltage-activated Kv3, Kv2 as well as Na v activities and exclude HCN or Kv4 (A-current [Kv4.2, Kv4.3]) channels, and, in combination with specific pharmacology, separated Kv3 and Kv2 currents. We found that Kv2 potassium channels were suppressed via a cGMP-independent and reversible redoxsensitive pathway, suggesting the involvement of S-nitrosylation.
Sodium channel activities were suppressed by NO, requiring both, a cGMP-dependent and redox-sensitive component, indicating combined and possibly additive phosphorylation and S-nitrosylation mechanisms. Effects on both ion channels led to a widening of AP waveforms and reduction in firing rates of pyramidal neurons. 2.2 | In vitro brain slice preparations C57/BL6 mice aged P14-P22 were killed by decapitation and slices containing the hippocampus were prepared as previously described (Steinert et al., 2011). Data were generated from 28 female mice with 3-4 brain slices per mouse and 3-8 neurons per slice. Horizontal slices (200 μm thick) were cut in a low-sodium artificial CSF (aCSF) at $0 C. Slices were maintained in a normal aCSF at 37 C for 1 h, after which they were stored at room temperature ($20 C). Composition of the normal aCSF was (in mM) as follows: NaCl 125, KCl 2.5, NaHCO 3 26, glucose 10, NaH 2 PO 4 1.25, sodium pyruvate 2, myoinositol 3, CaCl 2 2, MgCl 2 1, and ascorbic acid 0.5; L-arginine 0.1; pH was 7.4, bubbled with 95% O 2 , 5% CO 2 . For the low-sodium aCSF, NaCl was replaced by 250 mM sucrose, and CaCl 2 and MgCl 2 concentrations were changed to 0.1 and 4 mM, respectively. Osmolarity was adjusted to 310 mosmol/L.

| Electrophysiology
Whole-cell patch recordings were made from pyramidal neurons visually identified in the mid-third CA1 region in acute brain slices of the hippocampus as described previously (Bradley et al., 2017). Patch pipettes were pulled from glass capillaries (GC150F-7.5, o.d. 1.5 mm, Harvard Apparatus, Edenbridge, UK) with resistances of 3-5 MΩ.
Series resistances (R s ) were between 7 and 15 MΩ (R s compensation and prediction were set to 75-85%, electrode and whole-cell capacitance were compensated). Data were obtained using a Multiclamp 700B amplifier and Digidata 1550 ADDA converter. Acquisition (20 kHz sampling rate) and analysis were performed using pClamp 10.4 (Molecular Devices, Sunnyvale, CA). In current clamp, the amplifier circuitry was used to compensate the bridge balance (with pipette capacitance fully neutralized) and for experiments involving measurements of AP firing, current was injected to hold the membrane potential at À60 mV. AP firing threshold was defined as the point of maximum curvature of the voltage waveform between the onset of stimulation and the spike maximum at rheobase (Sekerli et al., 2004).
The pipette solution for whole-cell recordings contained 120 mM potassium gluconate, 10 mM HEPES, 0.2 mM EGTA, 4 mM K-ATP, 0.3 mM Na-GTP, 10 mM KCl, 8 mM NaCl, pH 7.4 with osmolarity set between 280 and 290 mosmol/L. All experiments were carried out at 33 ± 1 C using feedback control of a Peltier device warming the perfusing aCSF (1 ml/min).
In voltage-clamp, holding potentials were À60 mV and following a hyperpolarizing prepulse to À90 mV (360 ms) and additional prestep to À40 mV (35 ms), voltage command steps ranged from À110 to +50 mV (100 ms) to record voltage-activated potassium currents (Kv).
The prestep to À90 mV removed steady-state inactivation of Kv channels and the prestep to À40 mV induced inactivation of Kv4 currents and voltage-gated sodium currents to isolate high-voltage activate Kv2 and Kv3 conductances (Choudhury et al., 2020). The recorded potassium currents generated current-voltage (I-V) relationships (current measures were taken at a latency of 20 ms of the depolarizing step). Activation curves were fitted by the Boltzmann , where I is the tail current recorded at 0 mV after a prestep to command potential V (currents were estimated from the amplitudes of exponentials backfitted to the beginning of the test step), V half (V 1/2 ) is the membrane potential at which I is equal to onehalf I max and k is the slope factor. APs were evoked in current clamp by 12.5 pA-step current injections ranging from 0 to +200 pA for constructing current input-output relationships. All AP properties were determined from the first AP elicited at rheobase. To illustrate AP dynamics between groups, the changes in AP waveforms were demonstrated by generating phase plots. Input resistances (R in ) were determined according to Ohm's law by measuring the steadystate voltage deflection generated by 50 pA hyperpolarizing current injection for a 180 ms duration. Then, 50 pA of hyperpolarizing current was chosen because such a current intensity minimized the potential sag (Zemankovics et al., 2010). The membrane time constant (τ) was calculated by fitting a single exponential curve to the charging trajectory of the membrane potential between 10 and 90%, see Table 1 for passive membrane properties.
Sodium current steady-state voltage dependent inactivation was determined by a voltage protocol consisting of a 500 ms prepulse to potentials ranging from À90 to À15 mV (V, prepulse potential) followed by a 100 ms test pulse to À30 mV. Peak sodium currents were measured at a latency between 0.5 and 1.5 ms. The steady-state inactivation curves were constructed from currents activated at the corresponding prepulse potential (V) fitted to the Boltzmann function: . To minimize space clamp errors in nonspherical CA1 pyramidal neurons, we excluded recordings with R s > 10 MΩ and cell capacitance above 150 pF. In addition, we reduced current amplitudes by measuring at half-maximal channel activation (À30 mV) and excluded sodium current recordings which showed current activation with substantial delays, which would be a characteristic of the activation of channels in neurites (axon and dendrite) that are not adequately voltage-clamped (Cummins et al., 2009 clamping errors is to record smaller currents. As such, in Figure 1 we compared in addition to currents measured at +50 mV (~10 nA) also currents at 0 mV (~2-3 nA). Currents at both voltages revealed identical changes under the test conditions (NOC-5 application). We therefore focused in subsequent figures on Kv currents measured at +50 mV. All data points are from unpaired experiments and recorded at indicated time points of drug exposure with each data set having its own controls. Voltages were not corrected for liquid junction potentials.

| Statistical analysis
Results are expressed as the means ± SEM. Statistical comparisons were made by one-way ANOVA with Newman-Keuls posteriori analysis, two-way ANOVA with Bonferroni's multiple comparisons test, as appropriate. All data sets were tested for normality using the Shapiro-Wilk test showing normal data distributions. A p-value smaller than .05 was considered as significant. Asterisks directly above bars indicate p-values relative to corresponding controls, each data set has its own controls. Significance between other conditions is indicated by a line above the bars. All statistical tests were performed using Gra-phPad Prism 9 software.

| NO suppresses voltage-gated potassium currents
Divergent physiological phenotypes have been reported in response to NO applications or endogenous release due to the complexity of its downstream signaling. We decided to investigate the effects of NO in intact native pyramidal CA1 neurons allowing the required physiological signaling to occur. We first wanted to address how NO, delivered by a NO donor (NOC-5, 10 μM, equal to~400-500 nM of released NO (Bradley & Steinert, 2015)), affects voltage-gated potassium cur- 0013; between +10 and +50 mV@both timepoints: p < .0001, two-way ANOVA). Comparing outward currents at +50 mV shows that NO suppressed currents between 5 and 50 min (Figure 1(b), Ctrl vs. NOC: p = .0004 @5 min; p = .0051 @10 min, p = .0031 @20 min, p = .0003 @30 min, p = .0264 @50 min, one-way ANOVA). Passive intrinsic membrane properties were not affected by NO (Table 1).
Although R s values were indifferent between test conditions, larger currents can cause voltage clamping errors. In order to lessen those potential errors, we also compared smaller currents at 0 mV, typically around 2-3 nA, and found that NO similarly suppressed current amplitudes at this voltage following 20 min of exposure ( Figure 1(b), lower graph, Ctrl vs. NOC: p = .0101 @20 min; p = .0072 @30 min, p = .0223 @50 min, one-way ANOVA). Next, we wanted to elucidate the signaling routes by which NO exerts its effects. First, we exposed slices to NOC-5 in the presence of 10 μM ODQ (with 10 min preincubation) to block the activation of the endogenous NO receptor soluble guanylyl cyclase (sGC) that synthesizes cGMP from GTP (IC 50 for sGC:~1 μM (Garthwaite et al., 1995)). Inhibition of sGC did not These data show that currents recorded from neurons which were exposed for 50 min to NO followed by a 25 min NO-free aCSF washout remained unaltered, suggesting a reversible current modulation ( Figure 1(d), p = .5521 @washout, one-way ANOVA). To further support the notion of a cGMP-independent mechanism, we specifically confirmed that pharmacological activation of this pathway per se had no effect on outward currents. We recorded currents following incubation with an activator of sGC (YC-1) between 5 and 30 min (Wu et al., 1995), while simultaneously blocking the degradation of cGMP by phosphodiesterases (IBMX) resulting in maximally activated cGMP production (Friebe et al., 1998). Activation of sGC-cGMP signaling did not alter current amplitudes at any time point tested F I G U R E 1 Nitric oxide (NO)-mediated suppression of high-voltage activated potassium currents in CA1 pyramidal neurons is cGMPindependent but redox sensitive. (a) Left, sample voltage-clamp recordings from a control and NO-treated neuron (10 μM NOC-5) with voltage protocol below (arrow indicates latency at which current was measured). Right, summary IV curves for two timepoints: NO treatment reduced outward currents after 5 min at voltages positive to +10 mV and after 30 min at voltages positive to 0 mV (5 min NOC vs. Ctrl and 30 min NOC vs. Ctrl, two-way ANOVA). (b) Summary of mean current amplitudes at +50 and 0 mV. (c) Left, sample recordings from a control and NOC + ODQ-treated neuron. Right, summary IV curves for control and treated neurons: NOC + ODQ treatment suppresses outward currents between 20 and 50 mV (wash out vs. Ctrl: p > .05 and wash out vs. NOC + ODQ at 40 and 50 mV: p < .05). (d) Summary of mean current amplitudes at +50 mV. (e) Summary of mean current amplitudes at 50 mV at indicated timepoints following incubation with YC-1 (10 μM) + IBMX (10 μM), 5-50 min NOC data are average data taken from (c) for comparison. (f) Summary of mean current amplitudes at 50 mV at indicated timepoints following incubation with 8-Br-cGMP (10 μM), 5-50 min NOC data are average data taken from (c). (g) Summary of mean current amplitudes at 50 mV at indicated timepoints following incubation with NOC-5 (10 μM) + glutathione (GSH, 1 mM), 5-50 min NOC data are average data taken from (c) for comparison. Data denote mean ± SEM (n = 6-14 cells). Significance tested by two-way analysis of variance (ANOVA) (a,c) and one-way ANOVA (b,d,e-g) with *p < .05, **p < .01, ***p < .001, ****p < .0001 confirming that cGMP signaling is not involved in the observed nitrergic effects. Complementary experiments to activate PKG directly by 10 μM 8-Br-cGMP, a nonhydrolysable analogue of cGMP (Lu & Hawkins, 2002), also failed to alter whole-cell currents (Figure 1(f We concluded that cGMP and PKG activation are unlikely to be responsible for the nitrergic effects on outward currents and that the mechanism to regulate ion channel protein function may be via NO-induced reversible S-nitrosylation. We next tested whether suppressing the redox-sensitive S-nitrosylation impacts on the effects caused by NO by incubating the slices with membrane permeable GSH prior to (20 min) and during NO exposure. GSH has a strong reducing capacity and is the main cellular reactant to prevent cysteine S-nitrosylation (Clementi et al., 1998;Stomberski et al., 2019). Incubation  (Du et al., 1996;Martina et al., 1998). To determine which of the channels is affected by nitrergic GSH-dependent modulation, we used the Kv3 channel blocker, TEA, at a concentration of 1 mM to specifically eliminate Kv3 (Coetzee et al., 1999). Incubation with 1 mM TEA blocked~10% of total outward currents only at +50 mV (c) Summary of mean current amplitudes at 50 mV at 30 min exposure. Data denote mean ± SEM (n = 8-18 cells). Significance tested by two-way ANOVA (a,b) and one-way ANOVA (c) with *p < .05, **p < .01, ****p < .0001 If NO modulates a specific subset of outward currents we would predict that half-activation voltages (V 1/2 ) of the remaining currents may change, as different voltage-gated currents differ in their activation voltage range (Du et al., 2000;Lien et al., 2002;Martina et al., 1998).
Furthermore, activation kinetics change in response to differential channel clustering and posttranslational modifications as reported for Kv2 (Du et al., 2000;Misonou et al., 2004;Mohapatra et al., 2009). To determine V 1/2 values we measured tail current amplitudes over an activation range from À70 to +70 mV (Figure 3(a)) and fitted a Boltzmann function to the relative current amplitudes. Fitted curves for controls and following NO application are shown in Figure 3(b) (top). V 1/2 values in control neurons were in the range of 0 to +5 mV (Figure 3(b), bottom) as reported previously for pyramidal CA1 neurons (Mohapatra et al., 2009). Interestingly, incubation with NOC-5 caused a hyperpolarizing shift of V 1/2 to values of around À10 mV (Figure 3(b), bottom, one-way ANOVA). This shift was completely abolished by GSH and unaffected by ODQ (Figure 3(c,d), one-way ANOVA) suggesting that the nitrergic effects on V 1/2 were directly mediated by NO. Importantly, following NO exposure in the presence of TEA, we saw a hyperpolarizing shift in V 1/2 relative to TEA alone, and despite being of smaller magnitude (Figure 4(a), one-way ANOVA), the data confirm a modulation of high-voltage-activated currents other than Kv3. However, when blocking Kv2 by GxTx-1E, the NO-induced changes of V 1/2 were not observed (Figure 4(

| NO induces wider APs and reduces intrinsic excitability
Having established that NO modulates Kv2 currents, we next wanted to assess the effects of NO on the basic elements of information transmission including the characteristics of an AP and associated neuronal excitabilities. As high voltage-activate potassium currents, in addition to voltage-gated sodium and low voltage-activated potassium currents shape an AP waveform and set firing thresholds, we assessed these different AP parameters (half-width, amplitude and threshold) under various conditions of nitrergic activity. We expected an increase in half-width since the repolarizing phase of an AP in the hippocampus is determined by the strength of Kv2 currents (Liu & Bean, 2014). Indeed, we found that NO exposure over the time course of 5-50 min resulted in a substantial increase in half-widths of depolarization-evoked APs ( Figure 5(a,b), half-width of first AP at rheobase: Ctrl vs. NOC: 5 min: p = .0207, 10 min: p = .255, 20 min: p = .017, 30 min: p = .0172, 50 min: p = .0262, one-way ANOVA).
This AP widening will induce a significant effect during prolonged Although cGMP signaling itself did not impact on Kv currents, we tested whether AP waveforms were affected by pharmacological modulation of cGMP production. We confirmed that NO exposure in the presence of ODQ led to an increase in AP half-widths (first AP at rheobase), again confirming a cGMP-independent mechanism (  Significance tested by one-way analysis of variance (ANOVA) with ***p < .001, ****p < .0001 Together, these data imply two different effects induced by NO. First, an AP widening caused by a reduction in Kv2 currents which was unaffected by ODQ. Second, we noticed a NO-mediated suppression of AP firing rates during the 180 ms depolarization pulse.
We further noticed that amplitudes of the initial AP were reduced following NO exposure. Finally, we showed that the maximal dV m /dt values at the AP rising phase were substantially reduced by NO. These data point toward an additional effect of NO on sodium channel activities. Interestingly, the NO effects on AP amplitudes, firing rates and upstroke values were partially prevented by ODQ ( Figure 5(d),f,g,j)).
AP thresholds were unaffected by ODQ treatment (Figure 5(k), p > .05, one-way ANOVA). 8-Br-cGMP did not affect any of the AP parameters nor did direct stimulation of the sGC while blocking phosphodiesterase activities (Figure 6; 8-Br-cGMP and YC-1 + IBMX, p > .05, one-way ANOVA) with neither condition altering the firing rates during current-evoked depolarization (not shown). Since GSH impacted on Kv current modulation following NO exposure, we next tested whether GSH was also able to recover APs waveforms and firing pattern. Interestingly, GSH completely blocked the effects of NO on AP half-width, again confirming a redox-sensitive component of NO signaling (Figure 7(a,b)).
The lack of AP widening also impacted on the excitability, with GSH preventing the decline in firing rates during depolarizing current injections in the presence of NO (Figure 7(c)). When assessing the AP waveform, the phase plots illustrate the actions of GSH on nitrergic changes (Figure 7(d)) with GSH preventing the NOinduced reduction in AP amplitudes (Figure 7(e,f )). one-way ANOVA). In contrast, GxTx-1 increased AP half-widths bỹ 25% as reported previously (Liu & Bean, 2014) with NO having no additional effects (Figure 8(a) [-Kv2], Ctrl vs. GxTx-1: p = .008, Ctrl vs.
TEA + NOC: p = .0086, one-way ANOVA). Interestingly, blocking either Kv channel does not reduce AP numbers in responses to short current-induced depolarization as reported previously in the hippocampus following GxTx-1 application (Liu & Bean, 2014). However, additional NO application diminished the number of APs fired in the presence of TEA (at and above 87.5 pA current injection) and GxTx-1 (at and above 75 pA current injection, black arrows, TEA: p = .0133, GxTx-1: p = .0229, Figure 8( that inhibition of Kv2 in hippocampal neurons had no effect on initial AP firing rates but the observed changes in the current input-output relationships following NO exposure strongly imply that an additional mechanism, for example, reduced sodium channel activity, could be the underlying reason for the reductions in AP numbers.

| NO suppresses voltage-gated sodium currents in a cGMP-dependent and redox-sensitive manner
The observed decrease in AP amplitudes and a Kv-independent reduction in depolarization-evoked AP firing following NO exposure strongly indicates a modulation of voltage-gated sodium currents (Renganathan et al., 2001). We therefore wanted to test whether voltage-gated sodium channels were directly affected by NO. In baroreceptor neurons and N1E-115 neuroblastoma cells, previous studies have shown that application of NO donors can reduce voltage-gated sodium currents (Ribeiro et al., 2007); however, such data in hippocampal neurons has not yet been reported. At this point, we need to be cautious of voltage and space clamp errors which will impact on measurements of fast and large voltage-gated current recordings in nonspherical neurons (Williams & Mitchell, 2008). However, Na V current recordings have been reported in CA1 pyramidal neurons in acute brain slices (French et al., 2016;Liu et al., 2017) and applying rigorous criteria to minimize these potential errors, see Methods, enabled us to detect nitrergic effects on current amplitudes. We measured sodium currents activated by a step depolarization from À90 to À30 mV and found that NO reduced current amplitudes following 5-50 min of exposure by about 50% (Figure 9(a-c), one-way ANOVA). To characterize the potential nitrergic signaling routes, we assessed the effects of GSH and ODQ by co-incubating the slices with either drug. In parallel to the observed effects on potassium currents, we found that GSH prevented NO-induced current suppression (Figure 9(d-f)), suggesting that the inhibition of sodium channels might occur via redox signaling. To the contrary to the effects on Kv currents, inhibition of cGMP generation by ODQ also blocked the nitrergic effects (Figure 9(g-I)) indicating that both, cGMP-dependent and -independent nitrergic signaling were required to suppress Na V currents. This finding also agrees with data obtained on AP amplitudes and upstroke velocities (dV m /dt) where their reductions by NO were also prevented by both, GSH and ODQ ( Figure 5(g-j)). However, when specifically activating the cGMP cascade directly by 8-Br-cGMP or YC-1 + IBMX, both in the absence of NO, we did not observe any effects on sodium currents (Figure 10, p > .05, F I G U R E 6 Activation of cGMP signaling does not affect action potential waveforms. Summary data of action potential (AP) half-width, amplitude, and threshold following application of 8-Br-cGMP (a) and YC + IBMX (b). Data denote mean ± SEM (n = 6-18 cells). Significance tested by one-way analysis of variance (ANOVA), p > .05 one-way ANOVA) indicating that cGMP signaling alone is not sufficient to suppress the current.
Accumulating sodium current steady-state inactivation is one of the reasons for reduced AP amplitudes during high-frequency trains (Jung et al., 1997) and may mediate the reduction in firing rates observed in the presence of NO. We thus wanted to test whether NO affected channel inactivation per se by applying conditioning predepolarizing steps ranging from À90 to À15 mV for 350 ms from a holding potential of À90 mV followed by current activation at À30 mV (modified protocol from Sole et al. (2020)). Boltzmann fit to the relative current amplitudes in control conditions and following NO exposure revealed identical V 1/2 values, similar to previously reported values for channel inactivation in CA1 pyramidal neurons (Figure 11(a,b), p > .05, one-way ANOVA) (French et al., 2016). Furthermore, neither direct activation of the cGMP signaling nor addition of GSH (+NO) revealed any changes in V 1/2 or slope values (Figure 11(a,b), k Ctrl : À3.1 ± 0.3, k NOC : À4.8 ± 1.8, k NOC + GSH : À3.2 ± 0.6, k YC-1 + IBMX : À2.5 ± 0.6, one-way ANOVA). The data show that GSH or ODQ alone were able to prevent NO effects on sodium current amplitudes which suggests that the channel inhibition requires both pathways to be simultaneously active.
Together, our data illustrate that NO suppresses Kv2 currents in a redox-sensitive manner which translates into wider APs and a reduction in AP firing. However, an additional cGMP-and redox-sensitive component of NO signaling suppresses Na V currents which reduces AP amplitudes and firing rates independently of, and in addition to, Kv modulation.

| DISCUSSION
Our study revealed that NO differentially modulates hippocampal ion channel activities, thereby impacting on neuronal excitabilities and AP waveforms. We provide data illustrating nitrergic suppression of Kv2 channels in intact CA1 pyramidal neurons in a cGMP-independent manner likely due to NO-mediated posttranslational modifications. In addition, we show that NO reduced Na V currents, however, this NO-mediated effect required simultaneous cGMP-dependent and cGMP-independent redox signaling.
We confirmed that Kv2 current suppression was reversible and prevented by GSH, the most powerful endogenous denitrosylating conductance (Coetzee et al., 1999;Jan & Jan, 2012). However, future studies will have to specifically address the question whether Kv7 and Kv1 channels can be modulated by NO in hippocampal CA1 pyramidal neurons which could cause phenotypes of weaker and more modulatory nature.
Hippocampal Kv2.1 channels can undergo PKA-dependent phosphorylation (Ikematsu et al., 2011;Mohapatra et al., 2007), and our data suggest an additional regulatory pathway whereby NO, in a cGMP-independent manner, suppresses Kv2 currents in pyramidal neurons. Interestingly, pharmacological block of Kv2 in CA1 pyramidal neurons only leads to a decrease in AP firing following prolonged depolarizing steps >800 ms (Palacio et al., 2017). In contrast, our data indicate that NO reduces neuronal firing rates after only 100 ms of depolarization ( Figure 5(c)) which cannot be attributed to a sole nitrergic effect on Kv2 currents but rather suggests that an additional suppression of sodium currents contributes to the reduction in AP firing rates. As illustrated in the phase plots, NO strongly reduces the maximal upstroke of the AP (dV m /dt, Figure 5(g)) implicating a suppression of Na V (Liu & Bean, 2014). Our data show that, only when both, Kv2 and Na V conductances, are reduced, as would happen following endogenous NO generation, the modulatory action of NO results in complex effects on intrinsic excitability. neurodegeneration (Bourgognon et al., 2021;Dias et al., 2016;Du et al., 2020;Garthwaite, 2016;Pigott & Garthwaite, 2016;. Due to its nature as a volume transmitter, NO acts not only within the neuron of origin but also in neighboring cells, including astrocytes or microvasculature endothelium, a F I G U R E 9 Nitric oxide (NO) suppresses voltage-gated sodium currents in a glutathione-and ODQ-dependent manner. (a) Stepwise depolarizations of pyramidal neurons from a holding potential of À90 mV to test potentials between À90 and À15 mV elicited inward sodium currents recorded at À30 mV, see voltage protocol below. (b) Current-voltage relationship for peak current amplitudes at different times of NO incubation. (c) Summary data of peak amplitudes at indicated timepoints measured at maximal depolarization (from À90 to À30 mV). (d) Sample traces of stepwise depolarizations of pyramidal neurons from a holding potential of À90 mV to test potentials between À90 and À15 mV in the presence of NOC + glutathione ethyl-ester (GSH) (d) or NOC + ODQ (g). (e,h) Current-voltage relationship for peak current amplitudes at different times of co-incubation with GSH (e) and ODQ (h). (f,i) Summary data of peak amplitudes at indicated timepoints after co-incubation with NOC + GSH (f) and NOC + QDQ (i). Data denote mean ± SEM (n = 5-13 cells). Significance tested by one-way analysis of variance (ANOVA) with *p < .05, **p < .01 F I G U R E 1 0 Activation of cGMP signaling alone does not affect sodium currents. (a,c) Current-voltage relationship for peak current amplitudes at different times following 8-Br-cGMP (a) and YC + IBMX (c) incubation. (b,d) Summary data of peak amplitudes at indicated timepoints after incubation with 8-Br-cGMP (b) and YC + IBMX (d). Data denote mean ± SEM (n = 5-13 cells). Significance tested by one-way analysis of variance (ANOVA), p > .05 F I G U R E 1 1 Nitrergic effects on sodium currents are not due to changes in channel inactivation. (a) Boltzmann fitting to voltage-dependent inactivation of relative current amplitudes for indicated conditions. (b) Summary of mean V 1/2 values. (c) Summary of mean current amplitudes, averaged over 10-30 min for each condition. Data denote mean ± SEM (n = 8-19 cells). Significance tested by one-way analysis of variance (ANOVA) with *p < .05, **p < .01 mechanism related to neurovascular coupling, where its effects occur in a diffusion-limited space (Artinian et al., 2010;Garthwaite, 2019;Hoiland et al., 2020;Ledo et al., 2015;Lourenco et al., 2017;Philippides et al., 2005).
The multiple actions of NO impacting on the synaptic release machinery and ion channels may be mediated by the canonical cGMPmediated pathway in addition to specific posttranslational protein modifications (Bradley & Steinert, 2016;Gamper & Ooi, 2015;Robinson et al., 2018). However, the outcome of these processes depends on the rate of production and inactivation of reactive NO molecules  (Steinert, Postlethwaite, et al., 2010)) suggesting again an age-dependent decline in NMADR-mediated NO generation.
The volume diffusion of NO can occur with a radius of several tens of micrometers Philippides et al., 2005) and at concentrations as low as fmolar to pmolar whilst lasting several seconds at these levels (Batchelor et al., 2010;Lourenco, Ferreira, et al., 2014), indicating that volume transmitting has the ability to induce signaling within larger brain regions when released by a single neuron. The release profile of endogenously produced NO in response to NMDA perfusion into rat CA1 hippocampal regions shows a transient increase in NO levels between 100 and 350 nM (10-100 μM NMDA) (Ledo et al., 2005;Ledo et al., 2010). Importantly, different NO release profiles between the hippocampal CA1/CA3/dentate gyrus regions have been detected following transient glutamate injection into hippocampal brain slices from 8 to 10 weeks old rats (Lourenco, Ferreira, et al., 2014). The amounts of released NO vary drastically between the regions with peak levels reaching 3-4 fmol s À1 , lasting several tens of seconds.
Other reports show NO levels of several tens up to 300 pmol released in the cerebellum or hippocampus of P8-P10 old rats (Wood et al., 2011)-again, note that a younger age correlates with increased NO release. Due to the improvements of recording methods over the years, it is now possible to detect physiologically relevant NO concentrations although due to different methodologies (optical vs. electrochemical) and differences in animal ages, data comparisons are still difficult.
Due to the differences in NOS expression and NO generation between neurons and brain regions, reported physiological in vivo effects of NO signaling also vary greatly. In the auditory system, in vivo NO generation suppresses sound-evoked firing without affected basal activities (Hockley et al., 2020). Conversely, inhibition of NOS prevents the nicotine-induced increase in firing activity in neurons of the substantia nigra pars compacta without affecting their basal activities (Di Matteo et al., 2010). As these data reflect extracellular field recordings, it is difficult to conclude how NO acts on the level of ion channels. The vast amount of available and seemingly inconsistent data on NO signaling, further complicated by discrepancies between in vivo and in vitro conditions, makes this signaling route to one of the most studied, yet large gaps in our understanding remain. Our data offer new insights how NO can modulate intrinsic excitabilities in the hippocampus which potentially links activitydependent NO generation to changes in activities of ion channels, such as Na V and Kv2, responsible for membrane depolarization and hyperpolarization, respectively.
Our findings might be able to explain discrepancies between data obtained under in vivo and in vitro conditions. It has been long established that neuronal properties differ between the artificial slice preparations and in vivo environments. A recent study of a direct comparison of pyramidal neurons of the Layer 2/3 somatosensory cortex revelated that APs are broader and current-evoked firing frequencies are lower in vivo compared to in vitro (Fernandez et al., 2018). Although the authors only presented this observation and did not suggest any mechanisms, a likely relation to our findings could be an activity-dependent in vivo modulation of Kv2 and Na V currents by NO resulting in the observed differences.
The findings reported in our study may also have implications for regulating neuronal excitability following acute changes in hippocampal firing patterns and activity-driven NO generation, such as occur during traumatic brain injuries, epileptiform or changes in temporal spiking properties during fundamental computational activities (Ding et al., 2020;Ghotbeddin et al., 2019;Karimi et al., 2020;Martin et al., 2019;Song et al., 2020).
The use of NO donors provides an essential tool to study nitrergic effects, although unphysiologically high NO levels being released by various donors, pathological and disease-relevant iNOS-mediated NO release will be substantially higher and thus compare to higher NO donor release profiles. Numerous ion channel targets for NO signaling have been reported in in vitro preparations which provides detailed information on NO's ability to regulate neuronal excitability.
In Helisoma trivolvis snail B19 neurons, NO (100 μM NOC-7, 100 μM DEA-NONOate) inhibits SK (small calcium-activated potassium channel) as well as voltage-gated potassium channels but endogenous NO production enhances calcium and persistent sodium currents thereby adjusting neuronal excitability (Artinian et al., 2010;Artinian et al., 2012;Estes et al., 2015;Zhong et al., 2013Zhong et al., , 2015. NO inhibits voltage-gated and persistent sodium currents in C-type DRG neurons Future studies will have to identify whether the modulation of Na V channels by NO occurs directly on the channel protein or indirectly via other forms of posttranslational modifications such as phosphorylation or S-nitrosylation. Although it will have to be elucidated which exact posttranslational protein modifications occur on both ion channels, our data support the notion that activity-driven NO production might act as a feedback regulator to adjust ion channel activities and thus firing rates within in vivo networks.