Endoplasmic Reticulum and Mitochondrial Calcium Handling Dynamically Shape Slow Afterhyperpolarizations in Vasopressin Magnocellular Neurons

Abstract

Many neurons including vasopressin (VP) magnocellular neurosecretory cells (MNCs) of the hypothalamic supraoptic nucleus (SON) generate afterhyperpolarizations (AHPs) during spiking to slow firing, a phenomenon known as spike frequency adaptation. The AHP is underlain by Ca²⁺-activated K⁺ currents, and while slow component (sAHP) features are well described, its mechanism remains poorly understood. Previous work demonstrated that Ca²⁺ influx through N-type Ca²⁺ channels is a primary source of sAHP activation in SON oxytocin neurons, but no obvious channel coupling was described for VP neurons. Given this, we tested the possibility of an intracellular source of sAHP activation, namely, the Ca²⁺-handling organelles endoplasmic reticulum (ER) and mitochondria in male and female Wistar rats. We demonstrate that ER Ca²⁺ depletion greatly inhibits sAHPs without a corresponding decrease in Ca²⁺ signal. Caffeine sensitized AHP activation by Ca²⁺. In contrast to ER, disabling mitochondria with CCCP or blocking mitochondria Ca²⁺ uniporters (MCUs) enhanced sAHP amplitude and duration, implicating mitochondria as a vital buffer for sAHP-activating Ca²⁺. Block of mitochondria Na⁺-dependent Ca²⁺ release via triphenylphosphonium (TPP⁺) failed to affect sAHPs, indicating that mitochondria Ca²⁺ does not contribute to sAHP activation. Together, our results suggests that ER Ca²⁺-induced Ca²⁺ release activates sAHPs and mitochondria shape the spatiotemporal trajectory of the sAHP via Ca²⁺ buffering in VP neurons. Overall, this implicates organelle Ca²⁺, and specifically ER–mitochondria-associated membrane contacts, as an important site of Ca²⁺ microdomain activity that regulates sAHP signaling pathways. Thus, this site plays a major role in influencing VP firing activity and systemic hormonal release.

Significance Statement

The slow afterhyperpolarization (sAHP) is mediated by a Ca²⁺-dependent K⁺ current. Despite its critical role in regulating neuronal spiking, the Ca²⁺-dependent mechanisms leading to its activation and spatiotemporal shape remains poorly understood. Here we show that in vasopressin (VP) neurons, dynamic interactions in Ca²⁺ handling between endoplasmic reticulum (ER) and mitochondria play a significant role in sAHP initiation (via ER Ca²⁺ release) and its spatiotemporal waveform (via mitochondrial Ca²⁺ uptake). Our results suggest that contact sites between ER and mitochondria represent Ca²⁺ microdomains critically involved in initiating the first steps of sAHP generation in VP neurons. Given that changes in the sAHP have been linked to abnormal firing activity in various diseases, our results have both wide-range physiological and pathological implications.

Introduction

Magnocellular neurosecretory cells (MNCs) display different Ca²⁺-dependent hyperpolarizing afterpotentials that lower the intrinsic excitability of these neurons at different orders of magnitude of activation and duration such as the hyperpolarized afterpotential (HAP) and the afterhyperpolarization (AHP; Roper et al., 2003). The HAP is a hyperpolarizing overshoot of membrane potential lasting ∼25–125 ms and is sustained by an active conductance (Andrew and Dudek, 1984a; Bourque et al., 1985). In contrast, the AHP is an intrinsic excitability mechanism underlain by multiple Ca²⁺-dependent K⁺ currents that shape firing frequency on a longer timescale, shaping the patterning of spikes in many types of neurons (Andrade et al., 2012). It is usually described as having three components, underlain by AHP currents (I_AHP). The first is the fast AHP (fAHP) mediated by big conductance Ca²⁺-dependent K⁺ (BK) channels, which activate after single spikes and deactivate almost as quickly (Dopico et al., 1999). The second is the medium AHP (mAHP) which begins to appear after 3–5 spikes and has a decay time constant of ∼250 ms in VP neurons (Teruyama and Armstrong, 2005). The mAHP current (I_mAHP) is underlain by small conductance Ca²⁺-activated K⁺ (SK) channels and are sensitive to the bee venom apamin (Bourque and Brown, 1987; Schwindt et al., 1988a,b; Lorenzon and Foehring, 1992). The third component is the apamin-insensitive slow AHP (sAHP) underlain by a sAHP current (I_sAHP; Ghamari-Langroudi and Bourque, 2004; Villalobos et al., 2004; Armstrong et al., 2019). The sAHP duration lasts from seconds to tens of seconds and is Ca²⁺ dependent, is voltage independent, is insensitive to many K⁺ channel blockers, appears only after longer trains of spikes, and is strongly attenuated by neurotransmitters (Ghamari-Langroudi and Bourque, 2004; Andrade et al., 2012). The mechanisms for sAHP generation differ between neuronal cell types, and the ion channel(s) responsible remains elusive (Andrade et al., 2012). The sAHP is responsible for spike frequency adaptation and is impaired in many neurological diseases and conditions where neuronal firing is altered such as epilepsy and Alzheimer's disease (Moyer et al., 1992; McCormick and Contreras, 2001; Soh and Tzingounis, 2010; Kaczorowski et al., 2011). Thus, understanding sAHP molecular and ionic mechanisms is of critical physiological and clinical relevance.

In oxytocin (OT) and vasopressin (VP) MNCs of the supraoptic nucleus (SON), the sAHP activates during prolonged spike trains and plays a critical role in terminating the respective burst and phasic firing patterns that these neurons exhibit (Greffrath et al., 1998; Ghamari-Langroudi and Bourque, 2004; Teruyama et al., 2008). These patterns optimize neuropeptide release, minimizing fatigue by giving the neuron a cyclical refractory period (Dutton and Dyball, 1979; Bicknell and Leng, 1981; Bicknell, 1988). In VP neurons, the pattern is asynchronous between neurons, resulting in a smooth, sustained tone of VP release from the pituitary (Poulain and Wakerley, 1982; Bourque et al., 1998). As axonal hormone release tightly correlates with MNC activity, this mechanism consequentially shapes the pattern, frequency, and volume of hormones into the bloodstream from MNC axons terminating in the posterior pituitary (Bicknell and Leng, 1981; Cazalis et al., 1985; Armstrong et al., 1986; Gainer et al., 1986).

A rise in intracellular cytosolic Ca²⁺ concentration ([Ca²⁺]_c) is the first step in activating AHPs (Bourque et al., 1985; Lancaster and Adams, 1986; Ghamari-Langroudi and Bourque, 2004). While mAHPs are generated by direct Ca²⁺-gated SK activation (Adelman et al., 2012), sAHP mechanisms are slower and likely reflect recruitment of a signal transduction pathway (Vogalis et al., 2003; Villalobos et al., 2004; Andrade et al., 2012; Kim et al., 2016). The specific class of voltage-gated Ca²⁺ channels (VGCCs) contributing to AHP activation varies widely between neuronal cell types (Andrade et al., 2012). Previous work has demonstrated that N-type account for the majority of mAHP activation in OT and VP neurons (Kirchner et al., 2018). However, a clear cell-type discrepancy exists regarding activation of the sAHP between OT and VP neurons. While N-type VGCCs are the primary source of sAHP Ca²⁺ in OT neurons (Kirchner et al., 2018), the mechanism in VP neurons remains more complicated, as pharmacological block of any individual high voltage activated (HVA) VGCC failed to account for a majority of sAHP activation (Kirchner et al., 2018).

Three potential sources of activating Ca²⁺ remain untested. Either the sAHP reflects activation by low voltage activated T-type channels, a proportional rise in global [Ca²⁺]_c, or activation by an internal source of Ca²⁺ such as the endoplasmic reticulum (ER) or mitochondria. Given that the sAHP activates only after long spike trains, slow kinetics renders T-type Ca²⁺ an unlikely mechanism, as this channel has fast, transient biophysical properties (Erickson et al., 1993; Sato-Numata et al., 2020). Global [Ca²⁺]_c also seems unlikely given that previous work showed VGCC toxins significantly blocked (33–60%) [Ca²⁺]_c transients yet demonstrated no proportional block of the sAHP (Kirchner et al., 2018). The R-type blocker SNX-482 showed the greatest reduction, and the N-type blocker had only a minor yet statistically significant effect. Both N-type and R-Type blockers demonstrated an inconsistent corresponding reduction in Ca²⁺ amplitude. Thus, it appears that the sAHP time course does not simply reflect the trajectory of [Ca²⁺]_c. Studies have demonstrated that Ca²⁺ release from the ER activates the sAHP in other cell types (Sah and McLachlan, 1991; Torres et al., 1996; Shah and Haylett, 2000; Vogalis et al., 2001). Additionally, mitochondria can not only release Ca²⁺ to the cytosol (Takeuchi et al., 2015), but it also potently buffers [Ca²⁺]_c via the activity of the mitochondrial Ca²⁺ uniporter (Rizzuto et al., 2009, 2012), including in VP MNCs (Komori et al., 2010; Dayanithi et al., 2012). Furthermore, mitochondrial membranes form complexes with ER membranes (Rizzuto et al., 2009), including with RyRs (Szalai et al., 2000; Hajnóczky et al., 2002), and ER–mitochondria dynamic interactions have been shown to play an important role in regulating neuronal Ca²⁺ homeostasis (Rizzuto et al., 2009; Giorgi et al., 2015; Karagas and Venkatachalam, 2019). Given these collective findings, we tested here the role of these Ca²⁺-handling organelles in shaping sAHPs and firing properties in VP neurons.

Materials and Methods

Animals

All experiments were ethically approved by the Georgia State University Institutional Animal Care and Use Committee (IACUC) and performed in accordance with National Institutes of Health's Guide for the Care and Use of Laboratory Animals. All experiments were performed in 30 male and 24 female young adult Wistar rats in the late peripubertal to adolescent developmental stage (150–270 g) with transgenic expression of eGFP tagged to VP (Ueta et al., 2005), which were pooled for analysis. These rats received ad libitum food and water and housed on a standard 12 h light/dark cycle.

Ex vivo slice preparation

On the day of the experiment, rats were anesthetized with pentobarbital (50 mg kg⁻¹, i.p.) and then transcardially perfused with 30 ml of ice-cold sucrose artificial cerebrospinal fluid (aCSF) solution. This sucrose aCSF solution contained the following (in mM): 200 sucrose, 2.5 KCl, 1 MgSO₄, 26 NaHCO₃, 1.25 NaH₂PO₄, 20 d-glucose, 0.4 ascorbic acid, and 2.0 CaCl₂, pH 7.2, 300–305 mOsmol L⁻¹. The animal was then rapidly decapitated, and the brain was subsequently removed, mounted in the chamber of a vibratome (Leica VT1200S, Leica Microsystems) with superglue, and submerged into sucrose aCSF and bubbled constantly with 95% O₂/5% CO₂. Slices were cut at 240 µm thickness, bisected at the midline to separate the bilateral SONs into their own separate slice, and placed in a holding chamber containing standard aCSF (described below) bubbled with 95% O₂/5% CO₂. Slices rested in a water bath at 32°C for 20 min before transfer to room temperature for a minimum of 40 min before any recording.

Whole-cell patch clamp

External recording solution (aCSF) was the following (in mM): 119 NaCl, 2.5 KCl, 1 MgSO₄, 26 NaHCO₃, 1.25 NaH₂PO₄, 20 d-glucose, 0.4 ascorbic acid, and 2.0 CaCl₂, 2 Na⁺ pyruvate with 5 mM CsCl added to block the slow depolarizing afterpotential, which overlaps in time course with the AHP (Ghamari-Langroudi and Bourque, 1998; Armstrong et al., 2010). Slices were placed in the chamber of an upright microscope and perfused constantly by aCSF warmed to 32°C at a flow rate of 2–3 ml min⁻¹. All AHP recordings were performed in the presence of synaptic blockers 10 µM 6,7-dinitroquinoxaline-2,3-dione (DNQX), 40 µM 2R)-amino-5-phosphonovaleric acid (AP5), and 100 µM picrotoxin to block fast synaptic currents (Tocris Bioscience). Current-clamp and voltage-clamp recordings were acquired using a MultiClamp 700B and digitized with a Digidata 1550b (Molecular Devices). Electrodes were pulled using a Flaming Brown horizontal puller (Sutter Instruments) from borosilicate capillaries (2.5–5 MΩ) and filled with internal solution (in mM): 135 KMeSO₄, 8 KCl, 10 HEPES, 2 Mg-ATP, 0.3 Na-GTP, 6 phosphocreatine, as well as 0.05 fluo-5F for Ca²⁺ imaging, 7.2 pH [285–295 mosmol (kg H₂O)⁻¹]. Once we achieved whole-cell configuration, we allowed the cell to rest for 5 min to ensure stability of the patch and equilibration between internal solution and cytosol. We took baseline and experimental recordings in the presence of CsCl and synaptic blockers. In current clamp, we adjusted holding current to bring the cell to a resting membrane potential of −55 mV. We evoked AHPs using two different categories of protocol. We either used a 2.5 s duration depolarizing current step (0–150 pA) or a pulse train consisting of 30 spikes at 20 Hz with a 5 ms pulse width. Data was acquired at 10 kHz. In voltage clamp, we evaluated I_AHP by stepping from −55 to +10 mV at 20 Hz for 30 pulses (5 ms pulse width). Data was acquired at 2 kHz. Cells that displayed a shift in series resistance that exceeded 20 MΩ or a 25% change from the start of the recording were discarded. Experimental reagents include 3 µM thapsigargin (TG; Tocris Bioscience), 10 mM caffeine (Sigma-Aldrich), 1 µM carbonyl cyanide m-chlorophenyl hydrazone (CCCP; Sigma-Aldrich), 20 µM Ru360 (Ru360; EMD Millipore), and 20 µM triphenylphosphonium (TPP⁺; Sigma-Aldrich).

Calcium imaging

Ca²⁺ imaging was performed simultaneously with whole-cell patch-clamp using an Andor Dragonfly 200 spinning disk confocal microscope system (Andor Technology). Ca²⁺ was evaluated via fluo-5F delivered through the patch pipette (0.05 mM). Calcium fluorescence was obtained by exciting fluo-5F (or fluo-4 in uncaging experiments) at 488 nm. We measured fluorescence changes at an emission wavelength of 520 nm.

Photolytic calcium uncaging

Ca²⁺ uncaging experiments were performed simultaneously with patch clamp and Ca²⁺ imaging. The pipette internal solution was the same as described above with a few key differences. First, 2 mM DM-Nitrophen Ca²⁺ caging compound was added along with 0.8 mM CaCl₂ (40% occupancy). Second, 2 mM Na-ATP replaced 2 mM Mg-ATP as Mg²⁺ competes strongly with Ca²⁺ for DM-Nitrophen binding (Faas et al., 2005). Third, fluo-4 (K_d= 0.35 µM) was used as the Ca²⁺ indicator instead of fluo-5F (K_d= 2.3 µM), as fluo-5F is unable to compete with DM-Nitrophen with its low Ca²⁺ binding affinity. After achieving whole-cell configuration, we allowed internal solution to reach equilibrium with the inside of the cell (5 min minimum) before proceeding with experiments. We photolytically activated the DM-Nitrophen with a 50 ms pulse of 405 nm light delivered to the entire field. There was a 2 min minimum latency between trials to ensure membrane potential and [Ca²⁺]_c returned to baseline.

Experimental design and statistical analyses

In MNCs the fast AHP (fAHP) is a rapid event lasting <15 ms (Dopico et al., 1999) and was not evaluated herein. We measured AHP amplitude at two time points after a given stimulus. Since the I_mAHP decay tau is ∼500 ms in MCNs (Teruyama and Armstrong, 2005), we evaluated AHP (current clamp) and I_AHP (voltage clamp) amplitudes at 0.1 s (mAHP + sAHP) and 1 s (sAHP) from the end of the pulse depolarization or pulse train (Teruyama and Armstrong, 2005). This is to measure potential effects of the early phase of the sAHP (0.1 s after pulse termination) while also capturing a phase known to contain no mAHP contribution (1 s after pulse termination). Duration of the AHP in current clamp is the time from 0.1 s after the pulse to when the potential reaches baseline potential. Area of the AHP and I_AHP is the integrated area under the curve. Neuronal gain was calculated by fitting a linear regression to spike frequency data (f-I slope). Data was analyzed in ClampFit 11 (Molecular Devices) and statistical analysis was performed in GraphPad Prism (Dotmatics).

Fluorescent Ca²⁺ imaging data was background subtracted and measured at the soma excluding the nucleus. All Ca²⁺ imaging data are reported as percentage relative change in Ca²⁺ indicator fluorescence divided by baseline fluorescence calculated from a mean of points preceding the stimulus (%ΔF / F). Peak Ca²⁺ is the highest Ca²⁺ response point. Ca²⁺ amplitude is the height of the Ca²⁺ peak relative to baseline. Ca²⁺ area is the integrated area under the curve. Raw data was extracted using FIJI (NIH), conversion to %ΔF/F was done via MatLab script (MathWorks) and measurement of waveforms was done in either Igor Pro 9.0 (WaveMetrics) or ClampFit 11. Statistical analysis was performed in GraphPad Prism.

All data was tested for normality using a D'Agostino–Pearson omnibus test and homogeneity of variance. If the data met these conditions, an appropriate parametric test was performed. Otherwise, an equivalent nonparametric test was performed instead. Full statistical test and designs for each experiment can be found in the corresponding figure legend. n values represent the number of cells. No more than one cell per slice was evaluated. Histograms and plots were generated in GraphPad Prism. Example traces were generated by Igor Pro 9. Figures were created in Illustrator (Adobe).

Results

Depletion of ER Ca²⁺ stores abolishes sAHPs in VP neurons

Previous work studying HVA VGCC type contributions to AHPs in SON VP neurons demonstrated that while N-type channels coupled to the mAHP, no one channel gated the sAHP (Kirchner et al., 2018). Here we tested the contribution of ER Ca²⁺ as a source to sAHP activation, given its major role in amplifying Ca²⁺ signals via CICR (Verkhratsky and Shmigol, 1996). We incubated slices in thapsigargin (TG, 3 µM) for 1 h prior to recording to deplete ER Ca²⁺ stores. We generated AHPs with a pulse of 30 spikes at 20 Hz to control for spike count and Ca²⁺ influx. TG depletion significantly inhibited evoked sAHP amplitude, area, and duration in VP neurons, while leaving [Ca²⁺]_c peak and area unaffected in both current clamp (Fig. 1A,B) and voltage clamp (Fig. 1C,D). We considered whether biological sex might influence the TG effect on sAHPs, and thus we analyzed the data controlling for sex. TG blocked the sAHP in VP neurons in both males and females (Extended Data Fig. 1-1). Therefore, data collected from both sexes was pooled in each experiment throughout the rest of this study. sAHP recordings from putative OT (eGFP negative) neurons were unaffected by TG (Fig. 1), confirming the previous report that OT sAHPs are activated primarily by Ca²⁺ influx through N-type VGCCs (Kirchner et al., 2018). Furthermore, TG inhibition of sAHPs translated to meaningful changes in firing discharge, increasing firing rate in VP neurons. TG significantly increased evoked frequency and neuronal gain (f-I slope) in response to single depolarizing steps (Fig. 1E,F). These results suggest that ER Ca²⁺ gates the activation of sAHPs in VP neurons.

Figure 1.

Depletion of ER Ca²⁺ stores inhibits sAHPs in VP neurons. See Extended Data Figure 1-1 for an analysis of sex differences. A, Representative AHPs (arrows) recorded in current clamp (top) and corresponding Ca²⁺ responses (bottom) from four different groups: VP neurons under control conditions, VP neurons after 1 h (TG) incubation, OT neurons under control conditions, and OT neurons after 1 h of TG incubation. Note the inhibited AHP in the TG-incubated VP neuron (asterisk) but lack of effect in the OT neurons. B, Summary data of TG's effects on current-clamp AHPs evoked by a 20-spike 20 Hz depolarizing pulse train from −55 mV. Control (VP) n = 20, TG (VP) n = 40, control (OT) n = 14, TG (OT) n = 12. Statistics performed with a two-way ANOVA and Šídák's multiple comparisons test. p values for the overall test are provided in parentheses; for multiple comparisons, refer to the figure. 0.1 s amplitude (F = 8.49; p < 0.01), 1 s amplitude (F = 28.0; p < 0.0001), duration (F = 34.0; p < 0.0001), area (F = 30.3; p < 0.0001), Ca²⁺ amplitude (F = 4.04; p < 0.05), Ca²⁺ area (F = 0.17; p > 0.05). C, Representative I_AHP (arrows) recorded in voltage clamp (top) and corresponding Ca²⁺ responses (bottom) from four different groups: VP neurons under control conditions (left), VP neurons after 1 h TG incubation, OT neurons under control conditions, and OT neurons after 1 h of TG incubation. Note the inhibited AHP in the TG-incubated VP neuron (asterisk) but lack of effect in OT neurons. D, Summary data of TG's effects on voltage clamp I_AHP tail currents evoked by a 30-spike 20 s pulse train from a −55 mV holding potential. Control (VP) n = 13, TG (VP) n = 11, control (OT) n = 14, TG (OT) n = 6. Statistics performed with a one-way ANOVA and Šídák's multiple-comparisons test. p values for the overall test are provided in parentheses; for multiple comparisons, refer to the figure. 0.1 s amplitude (F = 30.6; p < 0.0001), 1 s amplitude (F = 15.7; p < 0.001), duration (F = 20.9; p < 0.0001), area (F = 24.0; p < 0.0001), Ca²⁺ amplitude (F = 1.08; p > 0.05), Ca²⁺ area (F = 2.96; p > 0.05). E, Representative spike trains evoked by a 70 pA depolarizing step under control conditions or after TG incubation. Note the increased spike frequency in the TG-treated neuron. F, Summary data of f-I relationships comparing VP neurons under control versus TG conditions. Control (VP) n = 5, TG (VP) n = 12. Statistics performed with a two-way ANOVA and Šídák's multiple-comparisons test. p values for the overall test are provided in parentheses; for multiple comparisons, refer to the figure. Spike count [F = 46.6; Control (VP) vs TG (VP) p < 0.0001), spike count (F = 51.8; Control (VP) vs TG (VP) p < 0.0001]. Neuronal gain (f-I slope) is calculated via linear regression significance [F = 5.5; Control (VP) 0.36 ± 0.04 vs TG (VP) 0. 19 ± 0.04; p < 0.05].

Figure 1-1

Thapsigargin incubation inhibits VP neuron AHPs similarly in male and female rats. AHP parameters of 1  h TG incubation in male and female rats. Male control n = 8; male TG n = 17; female control n = 12; female TG n = 23. Statistics performed with a two-way ANOVA and Šídák's multiple comparisons test. P-values for the overall group comparison are provided in parentheses, for multiple comparisons refer to the figure. 0.1  s amplitude (F = 10.26, p < 0.0001), 1  s amplitude (F = 13.00, p < 0.0001), duration (F = 10.73, p < 0.0001), area (F = 12.20, p < 0.01), Ca²⁺ peak (F = 6.49, p < 0.01), Ca²⁺ area (F = 4.38, p = 0.01). Download Figure 1-1, TIF file ^{(5.3MB, tif)} .

To further confirm an internal source of Ca²⁺ from the ER, we generated whole AHPs using photolytic uncaging of Ca²⁺ using DM-Nitrophen in VP neurons. This bypasses the need for Ca²⁺ entry via VGCCs during depolarization. As shown in Figure 2, Ca²⁺ uncaging efficiently evoked a large increase in [Ca²⁺]_c as well as a sAHP. TG preincubation still significantly blocked most of the sAHP component, despite robust Ca²⁺ increases in response to Ca²⁺ uncaging (Fig. 2A,B). It should be acknowledged that Ca²⁺ uncaging may activate other Ca²⁺ dependent currents such as Ca²⁺-dependent Cl⁻ channels (Paik et al., 2020), but we argue that contributions from other Ca²⁺-dependent phenomena are likely minimal due to similar time course of the uncaged response, consistent with sAHPs generated by current injection. Further, the response is inhibited by thapsigargin in a similar manner. The Ca²⁺ uncaging internal solution does differ significantly from the standard internal solution used in other experiments to accommodate for the properties of DM-Nitrophen. This notably includes the use of an alternative Ca²⁺ indicator with a different K_d (fluo-4). For these reasons, we retested action potential-evoked (30 spike 20 Hz spike trains) sAHPs and Δ[Ca²⁺]_c using the uncaging internal solution. Similar to our internal solution using fluo-5F, we found that cells preincubated with TG demonstrate significantly smaller sAHP components while Ca²⁺ transients remain unchanged compared with control conditions (Fig. 2C,D). These results demonstrate that large [Ca²⁺]_c increases can evoke sAHPs without activating VGCCs and that these sAHPs were significantly inhibited by depletion of ER stores.

Figure 2.

TG block of VP AHPs persists when AHPs are evoked by photolytic Ca²⁺ uncaging. A, Representative example of a Ca²⁺ trace (top) evoked by a single 50 ms pulse of 405 nm light (blue arrow) and the corresponding current clamp AHP evoked in response to the Ca²⁺ uncaging (bottom) under control conditions (left) and after TG incubation (right). B, Summary data of TG's effects on AHPs evoked by Ca²⁺ uncaging. Control n = 11, TG n = 8. 0.1 s amplitude (Mann–Whitney test; p < 0.5), 1 s amplitude (Mann–Whitney test; p < 0.05), duration (unpaired t test; t = 2.47; p < 0.05), area (unpaired t test; t = 2.27; p < 0.05), Ca²⁺ amplitude (unpaired t test; t = 0.01; p > 0.05), Ca²⁺ area (Mann–Whitney test; p > 0.05). C, Replication of current clamp AHPs evoked by 20 spike 20 Hz depolarizing pulse trains and corresponding Ca²⁺ responses using the modified Ca²⁺ uncaging internal solution containing Ca²⁺, a different Ca²⁺ indicator (fluo-4), and DM-Nitrophen which acts as a Ca²⁺ buffer. TG's effect on AHPs persists under these conditions. D, Summary data of TG's effects on AHPs evoked by 20-spike 20 Hz depolarizing pulse trains using the Ca²⁺ uncaging internal. Control n = 11, TG n = 8. 0.1 s amplitude (unpaired t test; t = 2.02; p > 0.5), 1 s amplitude (unpaired t test; t = 3.56; p < 0.01), duration (Mann–Whitney test; p < 0.0001), area (unpaired t test; t = 2.26; p < 0.05), Ca²⁺ amplitude (unpaired t test; t = 1.21; p > 0.05), Ca²⁺ area (unpaired t test; t = 1.60; p > 0.05).

ER Ca²⁺ release evoked by caffeine potentiates sAHPs in VP neurons

In addition to ER Ca²⁺ stores depletion, we assessed the role of ER Ca²⁺ release in sAHP activation by testing the effect of the ryanodine receptor (RyR) agonist caffeine (10 mM). Caffeine bath application failed to significantly increase the amplitude or the area of the whole AHP waveform generated by pulse trains (Fig. 3A,B). However, we observed a slight but statistically significant increase in the duration of the AHP. The Ca²⁺ peak was significantly smaller despite an increase in Ca²⁺ area (Fig. 3B). Additionally, measured input resistance showed a nonsignificant decrease in response to caffeine (control vs caffeine, 880.2 ± 197.1 vs 618.7 ± 126.6 MΩ; p = 0.063, Wilcoxon test). Interestingly, despite the small changes evoked by caffeine on the sAHP, we observed a significant decrease in spike frequency under this condition without a corresponding decrease in neuronal gain (f-I slope; Fig. 3C,D). Moreover, normalizing the amplitude of the AHP by the number of evoked spikes (AHP 0.1 s amplitude/spike count) revealed a significant increase in this AHP/spike index at lower depolarizing steps (10–50 pA) after caffeine application (Fig. 3D, right panel). Despite a general decrease in input resistance, the larger AHP per spike under the caffeine condition supports a genuine effect of caffeine on the AHP. Interestingly, Ca²⁺ amplitude decreases but cannot be attributed to saturation of the Ca²⁺ dye, as there is no indication that the signal plateaus. These results indicate that caffeine potentiates AHP activation via ER Ca²⁺ release following ryanodine receptor activation, leading to decreased firing rates (Fig. 3C).

Figure 3.

Caffeine bath application potentiates AHPs. A, Representative example of a current-clamp 30-spike 20 Hz AHP (top) and the corresponding Ca²⁺ responses (bottom) before and after caffeine. Current-clamp AHP evoked in response to the uncaging (bottom) under control conditions (left) and after TG incubation (right). B, Summary data of caffeine's effects on current clamp AHPs. Control and caffeine n = 8. 0.1 s amplitude (paired t test; t = 1.87; p > 0.5), 1 s amplitude (paired t test; t = 1.03; p > 0.05), duration (paired t test; t = 2.86; p < 0.05), area (Wilcoxon test; p > 0.05), Ca²⁺ peak (paired t test; t = 3.64; p < 0.01), Ca²⁺ area (paired t test; t = 2.52; p < 0.05). C, Representative spike trains evoked by a 70 pA depolarizing step before and after caffeine bath application. Note the decreased spike frequency in the neuron after caffeine. D, Summary data of AHPs before and after caffeine evoked by depolarizing steps. Control and caffeine n = 8. Statistics performed with a two-way ANOVA and Šídák's multiple-comparisons test. p values for the overall test are provided in parentheses; for multiple comparisons, refer to the figure. Spike Hz (F = 5.2; p < 0.5), AHP amplitude (F = 0.87; p > 0.5), AHP amplitude/spike count (F = 2.21; p = 0.05). Neuronal gain did not significantly differ between groups (control 0.16 ± 0.018 vs caffeine 0.15 ± 0.017; linear regression, F = 0.3, p > 0.05).

Blocking mitochondrial Ca²⁺ uptake enhances sAHPs

We next tested whether mitochondria, another critical organelle in intracellular Ca²⁺ handling, shapes the amplitude and time course of the sAHP here, as it has been observed in other cell types (Groten and MacVicar, 2022). Bath application of the mitochondrial uncoupler CCCP (1 µM), which disables mitochondrial via oxidative phosphorylation uncoupling and subsequent collapsing of the mitochondrial membrane potential (Miyazono et al., 2018), significantly enhanced all I_sAHP parameters along with corresponding increases in Ca²⁺ area (Fig. 4A,B). Interestingly, Ca²⁺ amplitude was slightly but consistently inhibited, suggesting a similar phenomenon to the caffeine effect wherein increased basal Ca²⁺ lowers the overall amplitude of the Ca²⁺ waveform. The overall enhancement of the Ca²⁺ area demonstrates that disabling mitochondria effectively restricted Ca²⁺ buffering. To further confirm that enhancements in I_sAHP reflected a lack of mitochondrial cytosolic Ca²⁺ buffering, we repeated this experiment with the more selective mitochondrial Ca²⁺ uniporter (MCU) blocker, Ru360 (20 µM) delivered through the patch pipette (Nathan and Wilson, 2017; Woods et al., 2019; Groten and MacVicar, 2022). After achieving whole-cell configuration, I_sAHP values were measured at 5, 10, and 15 min after break-in to track I_sAHP parameters in response to progressive Ru360 dialysis. Ru360 significantly enhanced I_sAHP amplitude, duration, and area over a 15 min recording period (Fig. 4C,D). Ca²⁺ amplitude consistently shrank over the 15 min, but this did not reach statistical significance, indicating once again that basal cytosolic Ca²⁺ is likely increasing. Ca²⁺ area significantly enhanced over this recording period, suggesting that Ca²⁺ buffering by mitochondria was impaired. Furthermore, subsequent bath application of CCCP (2 min) failed to further alter any of the I_sAHP or Ca²⁺ measurements (Fig. 4D), indicating that Ru360 dialysis occluded the CCCP effect, confirming in turn that the observed effect of CCCP alone (Fig. 4A) was due to Ca²⁺ buffering inhibition. Overall, these results indicate that mitochondria Ca²⁺ buffering normally restricts the I_sAHP time course in VP neurons, playing a significant role in shaping the AHP in these neurons.

Figure 4.

Blocking mitochondrial Ca²⁺ buffering enhances I_AHP. A, Representative voltage-clamp trace of I_AHP (left) and corresponding Ca²⁺ responses (right) before and after CCCP bath application. B, Summary data of CCCP's effects on I_AHP. Control and CCCP n = 7. 0.1 s amplitude (paired t test; t = 2.54; p < 0.5), 1 s amplitude (paired t test; t = 5.42; p < 0.01), duration (paired t test; t = 5.79; p < 0.01), area (Wilcoxon test; p < 0.05), Ca²⁺ peak (paired t test; t = 2.61; p < 0.05), Ca²⁺ area (paired t test; t = 4.40; p < 0.01). C, Representative voltage-clamp trace of I_AHP (left) and corresponding Ca²⁺ responses (right) during Ru360 perfusion through the patch pipette at 5, 10, and 15 min and then subsequent bath application of CCCP. D, Summary data of Ru360 + CCCP's effects on I_AHP. n = 10. Statistics performed with a repeated-measures one-way ANOVA and Šídák's multiple-comparisons test. p values for the overall test are provided in parentheses; for multiple comparisons, refer to the figure. 0.1 s amplitude (F = 11.4; p < 0.01), 1 s amplitude (F = 6.59; p < 0.05), duration (F = 20.51; p < 0.0001), area (F = 8.75; p < 0.01), Ca²⁺ peak (F = 5.27; p < 0.05), Ca²⁺ area (F = 9.35; p < 0.01).

Next, we determined whether blocking mitochondrial Ca²⁺ buffering affected sAHP time course in current-clamp mode and if so, if this blockade translated into slowing of repetitive firing properties. Dialysis of Ru360 through the patch pipette caused a significant enhancement of sAHPs generated by pulse trains over 9 min (Fig. 5A,B). This corresponded with a significant Ca²⁺ amplitude decrease and Ca²⁺ area enhancement, consistent with observations in voltage clamp (Fig. 4). Additionally, spiking during depolarizing steps significantly slowed after 9 min of Ru360 dialysis, compared with initial spike frequency measurements (Fig. 5C,D). These data suggest that not only do mitochondria play an important role in shaping AHP time course but that they efficiently influence repetitive firing frequency properties VP neurons.

Figure 5.

Blocking the mitochondrial Ca²⁺ uniporter with Ru360 enhances the AHP and slows firing in current clamp. A, Representative AHPs evoked with 30 spikes at 20 Hz (left) and corresponding Ca²⁺ responses (right) at 1 and 9 min after break-in. Note the enhancement of the AHP and Ca²⁺ signal time course after 9 min of perfusion. B, Summary data of Ru360 effects on AHPs compared with neurons without Ru360. Control n = 4, Ru360 n = 6. Statistics performed with a repeated-measures two-way ANOVA and Šídák's multiple-comparisons test. p values for the group comparison are provided in parentheses; for multiple comparisons, refer to the figure. 0.1 s amplitude (F = 28.72; p < 0.0001), 1 s amplitude (F = 14.52; p < 0.0001), duration (F = 18.86; p < 0.0001), area (F = 10.37; p < 0.0001), Ca²⁺ peak (F = 16.48; p < 0.0001), Ca²⁺ area (F = 13.69; p < 0.0001). C, Representative example of spiking in response to a 70 pA depolarizing step 1 min and 9 min after achieving whole-cell configuration with Ru360 in the pipette. Note the lower spike frequency at 9 min. Summary data of Ru360 effects on AHP spike frequency performed with a one-way ANOVA and Šídák's multiple-comparisons test (n = 5; F = 21.76; p = 0.001). For multiple comparisons refer to the figure.

Block of mitochondrial Ca²⁺ efflux does not significantly affect sAHPs

In addition to Ca²⁺ buffering via the MCU, mitochondria also contain pumps that can exude Ca²⁺ into the cytosol in a Na⁺-dependent or Na⁺-independent manner (Bernardi, 1999). Since mitochondrial buffering specifically modulates sAHP time course, we wanted to test whether blocking mitochondrial Ca²⁺ efflux would inhibit sAHPs. TPP⁺ is a compound that specifically inhibits mitochondrial Ca²⁺ efflux (Karadjov et al., 1986). It should be noted that TPP⁺ fully blocks Na⁺-dependent Ca²⁺ efflux while blocking ∼50% of Na⁺-independent Ca²⁺ efflux at the concentration used (10 µM; Wingrove and Gunter, 1986). Attempts at higher concentrations of TPP⁺ produced inconsistent results and often compromised cell viability in our preparation, so we used 10 µM despite incomplete block. Bath application of TPP⁺ failed to inhibit the AHP in either current clamp (Fig. 6A,B) or voltage clamp (Fig. 6C,D). Corresponding Ca²⁺ peak and areas after TPP⁺ application remained unchanged except that Ca²⁺ area increased in current clamp (Fig. 6B). Subsequent application of CCCP (1 µM) produced significant enhancements to AHP and I_AHP duration and area (Fig. 6B,D) as previously observed. Corresponding Ca²⁺ area, but not peak, significantly increased in both current and voltage clamp (Fig. 6B,D). Together, these results indicate that in stark contrast to mitochondrial Ca²⁺ uptake, mitochondrial Ca²⁺ efflux plays no consequential role in sAHP activation or modulation. To investigate the possibility that these compounds may be increasing the overall baseline Ca²⁺, we compared raw baseline Ca²⁺ signals from this data in both voltage and current clamp and found no significant differences between the three groups (Extended Data Fig. 6-1).

Figure 6.

Blocking mitochondrial Ca²⁺ pumps via TPP⁺ does not significantly effect AHPs. A, Representative example of TPP⁺'s effect on AHPs (top) and corresponding Ca²⁺ responses (bottom). Note that while TPP⁺ has little effect, subsequent application of CCCP enhances the AHP. See Extended Data Figure 6-1 for an analysis of baseline Ca²⁺ differences between the groups. B, Summary data of TPP⁺ effects on AHPs. n = 8. Statistics performed with a one-way repeated-measures ANOVA and Šídák's multiple-comparisons test. p values for the overall test are provided in parentheses; for multiple comparisons, refer to the figure. 0.1 s amplitude (F = 3.14; p > 0.05), 1 s amplitude (F = 6.18; p < 0. 05), duration (F = 22.51; p < 0.01), area (F = 8.01; p < 0.05), Ca²⁺ peak (F = 0.34; p > 0.05), Ca²⁺ area (F = 9.78; p < 0.05). C, Representative example of TPP⁺'s effect on I_AHP (top) and corresponding Ca²⁺ responses (bottom). Note that while TPP⁺ has little effect, subsequent application of CCCP enhances the AHP. D, Summary data of TPP⁺ effects on I_AHP. n = 11. Statistics performed with a one-way repeated-measures ANOVA and Šídák's multiple-comparisons test. p values for the overall test are provided in parentheses, for multiple comparisons refer to the figure. 0.1 s amplitude (F = 6.74; p < 0.05), 1 s amplitude (F = 6.32; p < 0.05), duration (F = 7.98; p < 0.05), area (F = 9.25; p < 0.01), Ca²⁺ peak (F = 3.79; p > 0.05), Ca²⁺ area (F = 10.38; p < 0.01).

Figure 6-1

Baseline Ca²⁺ measurements do not significantly differ after TPP⁺ or subsequent CCCP. Raw baseline fluorescence for Control, TPP⁺, and CCCP in both voltage clamp (left) and current clamp (right). In voltage clamp, we observed a significant difference (One-way repeated measures ANOVA, F = 4.80, p = 0.031) however post hoc analysis revealed no significant between group differences. In current clamp, we observed a significant difference (One-way repeated measures ANOVA, F = 4.11, p = 0.042) however post hoc analysis revealed no significant between group differences. Download Figure 6-1, TIF file ^{(2.3MB, tif)} .

Discussion

This study was designed to determine a Ca²⁺ source activating sAHPs in VP neurons. Previous work demonstrated N-type channels were primarily responsible for activating the mAHP and sAHP in OT neurons while coupling of specific VGCCs to the mAHP and sAHP in VP neurons was less clear (Kirchner et al., 2018). The largest contributions were a 29% reduction after N-type block via conotoxin GVIA and a 25% reduction in sAHPs after R-type block via SNX-482. Thus, no single Ca²⁺ source accounted for a majority of the sAHP in VP neurons, compared with a more comprehensive block of the OT AHP components by conotoxin GVIA (Kirchner et al., 2018). We therefore tested the contribution of ER Ca²⁺ stores to sAHP activation in VP neurons.

We show that depletion of ER stores via TG abolishes the sAHP, suggesting that ER Ca²⁺ plays a major role activating the sAHP in VP neurons (Fig. 1). Furthermore, RyR activation by caffeine appears to potentiate sAHP activation (Fig. 3) by a mechanism yet to be determined. This is similar to ryanodine-dependent sAHP activation described in other neuronal cell types (Pineda et al., 1999; Shah and Haylett, 2000; van de Vrede et al., 2007; Sahu et al., 2019) and overall suggests the sAHP in VP neurons relies heavily on Ca²⁺-induced ER Ca²⁺ release (CICR). One parsimonious explanation for the major contributors is that R-type and N-type channels provide Ca²⁺ that activates CICR from ER. Coupling between N-type VGCCs and ER in MNCs is previously established within the context of somatodendritic peptide release, a unique phenomenon of MNCs wherein Ca²⁺-dependent exocytosis of OT and VP occurs in the soma and dendrites (Ludwig, 1998; de Kock et al., 2003; Tobin et al., 2004; Brown et al., 2020). Thapsigargin depletion mobilizes large dense core vesicles (LDCVs) to readily releasable pools at plasma membrane (Ludwig et al., 2002; Tobin et al., 2004), and this mobilization is blunted in the presence of N-type blocker ω-conotoxin GVIA. Moreover, the N-type channel has high association to Golgi apparatus where peptide is packaged into LDCVs (Tobin et al., 2011). With regard to R-type VGCCs, close association with ER is less established but they interestingly show much higher expression in VP neurons compared with OT (Tobin et al., 2011). The significance of R-type channels in VP MNC neurophysiology requires further study.

The resultant [Ca²⁺]_c inhibition by VGCC blockade does not correspond with a proportional inhibition of the sAHP, suggesting that it is likely activated by a Ca²⁺ microdomain instead of reflecting the magnitude and temporal trajectory of a global [Ca²⁺]_c rise (regardless of the source). Evidence for this sort of proportionality exists in neocortical pyramidal neurons, where there is a sigmoidal cooperative relationship between I_sAHP amplitude and [Ca²⁺]_c (Abel et al., 2004). This does not appear to be the case in VP neurons, where we found that depletion of ER Ca²⁺ stores significantly reduces sAHPs (69–92%) without meaningfully reducing global [Ca²⁺]_c (Figs. 1, 2). This suggests that measurements of global [Ca²⁺]_c are a poor indicator of the properties of the evoked sAHP, arguing also against its contribution as an underlying mechanism triggering its activation. Global [Ca²⁺]_c increases are still critical to the activation of the sAHP, evidenced by its complete blockade when Δ[Ca²⁺]_c are prevented (Ghamari-Langroudi and Bourque, 2004; Kirchner et al., 2018). Rather, we propose that a Ca²⁺ threshold must be met at specific sites to activate relevant signaling cascades. It appears sAHP signaling activated by ER Ca²⁺ is mediated by a microdomain that is modulated without an obvious impact in global [Ca²⁺]_c. This is consistent with previous data in VP neurons showing that [Ca²⁺]_c decay is dissociated from AHP decay, suggesting compartmentalization of AHP-activating Ca²⁺ (Roper et al., 2003).

The significance underlying the contrasting sAHP Ca²⁺ sources of SON OT and VP neurons is at present unknown. Despite their indiscriminate morphological appearance, these neurons have considerably different electrophysiological properties (Armstrong and Stern, 1998; Armstrong et al., 2019), including differences in their dependence on PIP₂ to activate sAHPs and VGGCs (Kirchner et al., 2017, 2019). Thus, it is tempting to speculate that differences in sAHP Ca²⁺ sources could be associated to the differing physiological roles OT and VP play within the organism and their secretion in response to differing stimuli. An instructive example of this is the AHP plasticity that occurs in OT but not VP neurons during pregnancy and lactation, where it is hypothesized that this AHP enhancement helps facilitate the short, high-frequency OT milk ejection bursts (Stern and Armstrong, 1996; Teruyama and Armstrong, 2002, 2005; Wang et al., 2018). OT neurons can coordinate their activity and burst synchronously while VP neurons tend to demonstrate asynchronous phasic firing, which results in sustained systemic secretion of vasopressin needed to cope with an osmotic challenge (Bourque et al., 1998). Another notable difference between OT and VP neurons is their calbindin and calretinin expression levels. Previous studies report both these Ca²⁺-buffering proteins express in 70% of OT neurons while both are absent in 73% of VP neurons (Arai et al., 1999). This suggests that VP neurons have a lower overall Ca²⁺ buffering capacity compared with OT neurons, which could partially explain the role of CICR. Furthermore, this is consistent with reports that VP neurons generally have larger AHPs, though this difference is subtle (Armstrong et al., 1994; Stern and Armstrong, 1996; Teruyama and Armstrong, 2002). Clearly, future studies are needed to further understand the significance of the different Ca²⁺ sources triggering AHPs in VP and OT neurons.

We also tested the contribution of the mitochondria to sAHPs in VP neurons by either disabling them completely (Fig. 4), blocking the Ca²⁺ uptake uniporter MCU (Figs. 4, 5), or disabling the Na⁺–Ca²⁺ exchanger (Fig. 6). Blocking mitochondria Ca²⁺ uptake significantly enhanced the sAHP, while blocking Ca²⁺ extrusion had no effect, indicating that mitochondria does not contribute as a Ca²⁺ source to activate that sAHP but that it rather restricts the sAHP time course by actively buffering Ca²⁺. This is in line with a previous study showing that the primary Ca²⁺ handling role of mitochondria is to buffer it in VP neurons (Komori et al., 2010).

ER and mitochondria membranes can form tight complexes called mitochondria-associated membranes (MAMs; Rizzuto et al., 1998). These MAMs are dynamic protein scaffolds wherein high spatiotemporal control of local Ca²⁺ microdomains can coordinate signaling pathways (Patergnani et al., 2011; Giorgi et al., 2015; Missiroli et al., 2018). MAMs contain Ca²⁺ sensors, which are critically important for sAHP activation in other cell types, including hippocalcin in CA1 (Tzingounis et al., 2007), neurocalcin-δ cortical pyramidal neurons (Villalobos and Andrade, 2010), and calcineurin in myenteric neurons (Vogalis et al., 2004). MAMs may play an important role in providing a spatially constricted, microdomain avenue for sAHP activation in VP neurons. This is supported by our data showing a high efficiency of mitochondrial Ca²⁺ clearing combined with ER Ca²⁺ release to dictate the activation and spatiotemporal dynamics of the sAHP. These interactions are critical for not only understanding intrinsic excitability, but also VP somatodendritic release, as ER Ca²⁺ critically primes VP-containing LDCVs to the plasma membrane (Li and Hatton, 1997a; Tobin et al., 2004; Tobin and Ludwig, 2007; Kortus et al., 2016). In fact, given the strong overlap in Ca²⁺ sources by both the sAHP and somatodendritic release, it is tempting to speculate that sAHP may play a role in modulating the volume and pattern of somatodendritic release. Experiments to test this possibility as well as determining whether mitochondrial Ca²⁺ buffering plays a role in somatodendritic release are warranted.

The sAHP is one of many intrinsic excitability mechanisms that shape the phasic firing patterns of VP neurons (Thirouin and Bourque, 2021). During MNC burst onset, the reciprocal depolarizing afterpotential (DAP) activates, sustaining a plateau potential for phasic burst onset and regeneration via CICR activation (Andrew and Dudek, 1984b; Armstrong et al., 1994; Li and Hatton, 1997a,b). This present study suggests that the ER in VP neurons plays a role in regulating the intrinsic excitability of VP neurons by activating both DAP and sAHP, likely through a differing spatiotemporal pattern of ER Ca²⁺ release during the burst phase.

The sAHP is a mechanism that is commonly affected in many disease states where neuronal firing is abnormal. For example, cortical pyramidal neurons in epileptic subjects have diminished sAHPs, (Lorenzon and Foehring, 1992; Soh and Tzingounis, 2010; Tiwari et al., 2022) and neurons from Alzheimer's mouse models show enhanced sAHPs (Landfield and Pitler, 1984; Moyer et al., 1992; Power et al., 2002; Kaczorowski et al., 2011). Furthermore, previous work from our laboratory demonstrated hyperexcitability of VP neurons in a rat model of heart failure (Han et al., 2010; Potapenko et al., 2011, 2013; Stern and Potapenko, 2013; Ferreira-Neto et al., 2017; Ferreira-Neto and Stern, 2019), which contributes to neurohumoral activation, morbidity, and mortality in this prevalent disease (Goldsmith et al., 1983; Francis et al., 1990; Chatterjee, 2005). In this context, we demonstrated that a blunted mAHP current despite normal somatic Ca²⁺ elevation in response to pulse trains constitutes an underlying mechanism leading to VP hyperexcitability in this disease (Ferreira-Neto et al., 2017). Thus, future studies to determine whether the sAHP in VP neurons and the underlying ER CICR mechanism are also affected during heart failure are warranted.