Altered NMDA receptor‐evoked intracellular Ca²⁺ dynamics in magnocellular neurosecretory neurons of hypertensive rats

Abstract

Key points

• NMDA receptor (NMDAR)‐mediated Ca²⁺ signalling plays a critical role in modulating hypothalamic neurosecretory function. However, whether an altered NMDAR‐evoked changes in Ca²⁺ (NMDAR‐ΔCa²⁺) signalling in magnocellular neurosecretory cells (MNCs) may contribute to neurohumoral activation during disease states is unknown.

• We show that activation of NMDARs evoked similar inward currents in MNCs of sham and renovascular hypertensive (RVH) rats. Despite this, a prolonged and larger NMDAR‐ΔCa²⁺ response was observed in the latter.

• The exacerbated NMDAR‐ΔCa²⁺ responses in MNCs of RVH rats affected both somatic and dendritic compartments.

• Inhibition of the endoplasmic reticulum sarcoendoplasmic reticulum calcium trasport ATPase (SERCA) pump prolonged NMDAR‐ΔCa²⁺ responses in sham rats, but not in RVH rats.

• Our study supports an altered spatiotemporal dynamic of NMDAR‐ΔCa²⁺ signalling in MNCs from RVH rats, partly due to blunted endoplasmic reticulum Ca²⁺ buffering capacity.

Abstract

A growing body of evidence supports an elevated NMDA receptor (NMDAR)‐mediated glutamate excitatory function in the supraoptic nucleus and paraventricular nucleus of hypertensive rats that contributes to neurohumoral activation in this disease. However, the precise mechanisms underlying altered NMDAR signalling in hypertension remain to be elucidated. In this study, we performed simultaneous electrophysiology and fast confocal Ca²⁺ imaging to determine whether altered NMDAR‐mediated changes in intracellular Ca²⁺ levels (NMDAR‐ΔCa²⁺) occurred in hypothalamic magnocellular neurosecretory cells (MNCs) in renovascular hypertensive (RVH) rats. We found that despite evoking a similar excitatory inward current, activation of NMDARs resulted in a larger and prolonged ΔCa²⁺ in MNCs from RVH rats. Changes in NMDAR‐ΔCa²⁺ dynamics were observed both in somatic and dendritic compartments. Inhibition of the sarcoendoplasmic reticulum calcium trasport ATPase (SERCA) pump activity with thapsigargin prolonged NMDAR‐ΔCa²⁺ responses in MNCs of sham rats, but this effect was occluded in RVH rats, thus equalizing the magnitude and time course of the NMDA‐ΔCa²⁺ responses between the two experimental groups. Taken together, our results support (1) an exacerbated NMDAR‐ΔCa²⁺ response in somatodendritic compartments of MNCs of RVH rats, and (2) that a blunted ER Ca²⁺ buffering capacity contributes to the altered NMDAR‐ΔCa²⁺ dynamics in this condition. Thus, altered spatiotemporal dynamics of the NMDAR‐ΔCa²⁺ response stands as an underlying mechanism contributing to neurohumoral activation in neurogenic hypertension.

Key points

• NMDA receptor (NMDAR)‐mediated Ca²⁺ signalling plays a critical role in modulating hypothalamic neurosecretory function. However, whether an altered NMDAR‐evoked changes in Ca²⁺ (NMDAR‐ΔCa²⁺) signalling in magnocellular neurosecretory cells (MNCs) may contribute to neurohumoral activation during disease states is unknown.

• We show that activation of NMDARs evoked similar inward currents in MNCs of sham and renovascular hypertensive (RVH) rats. Despite this, a prolonged and larger NMDAR‐ΔCa²⁺ response was observed in the latter.

• The exacerbated NMDAR‐ΔCa²⁺ responses in MNCs of RVH rats affected both somatic and dendritic compartments.

• Inhibition of the endoplasmic reticulum sarcoendoplasmic reticulum calcium trasport ATPase (SERCA) pump prolonged NMDAR‐ΔCa²⁺ responses in sham rats, but not in RVH rats.

• Our study supports an altered spatiotemporal dynamic of NMDAR‐ΔCa²⁺ signalling in MNCs from RVH rats, partly due to blunted endoplasmic reticulum Ca²⁺ buffering capacity.

Abbreviations

eNMDAR

extrasynaptic NMDAR

ER

endoplasmic reticulum

MNC

magnocellular neurosecretory cell

NMDAR

NMDA receptor

OT

oxytocin

PVN

paraventricular nucleus

RVH rat

renovascular hypertensive rat

SERCA

sarco/endoplasmic reticulum Ca²⁺‐ATPase

SON

supraoptic nucleus

TG

thapsigargin

VP

vasopressin

Introduction

Vasopressin and oxytocin magnocellular neurosecretory cells (MNCs) of the hypothalamic supraoptic (SON) and paraventricular (PVN) nuclei play critical roles in fluid balance, and cardiovascular and reproductive homeostasis (Silverman & Zimmerman, 1983). Moreover, exacerbated activity of MNCs contributes to neurohumoral activation during cardiovascular diseases such as hypertension and heart failure (Riegger et al. 1985; Packer et al. 1987; Packer, 1988; Yemane et al. 2010; Littlejohn et al. 2013). Neurohumoral activation has a direct impact into morbidity/mortality in these diseases (Cohn et al. 1984; Yemane et al. 2010). Thus, elucidating the precise cellular mechanisms contributing to altered MNC activity in hypertension is of critical importance.

The amino acid glutamate, acting primarily on ionotropic NMDA receptors (NMDARs) is the major excitatory neurotransmitter within the SON and PVN (van den Pol et al. 1990). Binding of glutamate to NMDARs results in an influx of Ca²⁺ that evokes a direct membrane depolarization and the adoption of burst firing activity (Hu & Bourque, 1992; Nissen et al. 1995), which optimizes neuropeptide release from the posterior pituitary (Cazalis et al. 1985). Importantly, a growing body of evidence supports an elevated glutamate tone (Zhang et al. 2017), an increased glutamatergic innervation density (Biancardi et al. 2010) and increased expression of NMDARs (Li et al. 2014; Glass et al. 2015) in the SON and PVN of hypertensive rats.

In addition to evoking membrane depolarization, the NMDAR‐mediated Ca²⁺ influx results in an increase in intracellular free Ca²⁺ levels (ΔCa²⁺) (McBain & Mayer, 1994) affecting in turn a variety of downstream signalling mechanisms, including activation of Ca²⁺‐sensitive channels (Petersen, 2002; Sah & Faber, 2002), changes in the balance of kinase/phosphatase activities (Colbran & Brown, 2004) and stimulation of nitric oxide production (Bredt & Snyder, 1989), all of which could further alter MNC activity.

We recently showed that, in addition to acting on synaptically located NMDARs (sNMDARs) to mediate transient excitatory postsynaptic currents (EPSCs), glutamate also binds to extrasynaptic NMDARs (eNMDARs), activation of which mediates a tonic, sustained excitatory current that strongly stimulates firing activity in MNCs (Fleming et al. 2011). Moreover, we showed that eNMDARs (but not sNMDARs) are functionally coupled, in a Ca²⁺‐dependent manner, to the A‐type K⁺ channel‐mediated current I _A (Naskar & Stern, 2014) and that an augmented eNMDAR–I _A coupling is a critical mechanism contributing to exacerbated MNC activity in renovascular hypertension (RVH) (Zhang et al. 2017). However, whether a change in NMDAR‐ΔCa²⁺ contributed to the altered NMDAR–I _A coupling and increased MNC activity in RVH rats remains unknown.

The functional consequences of the NMDAR‐ΔCa²⁺ are largely dependent on the magnitude and time course of the Ca²⁺ signal, which in turn are determined not only by the amount of Ca²⁺ influx following NMDAR activation, but also by intracellular Ca²⁺ buffering mechanisms, including the endoplasmic reticulum via the sarco/endoplasmic reticulum Ca²⁺‐ATPase (SERCA) pump (Toescu & Verkhratsky, 1998). Surprisingly, the precise spatiotemporal Ca²⁺ dynamics in MNCs following NMDAR activation have not been investigated in detail, and more importantly, whether an altered NMDAR‐ΔCa²⁺ dynamic in MNCs occurs during hypertension, contributing in turn to its exacerbated actions, remains unknown. To this end, we performed in this study simultaneous patch‐clamp electrophysiological recordings and fast confocal Ca²⁺ imaging to characterize NMDAR‐ΔCa²⁺ spatiotemporal dynamics in MNCs in sham and RVH rats.

Methods

Ethical approval

All experimental procedures were in strict compliance with NIH guidelines, and were approved by the Augusta University Institutional Animal Care and Use Committee, and conform to the principles and regulations as described in Grundy (2015).

Animals

Male Wistar rats purchased from Harlan (Indianapolis, IN, USA, 4–5 weeks old) were housed under standardized conditions (12 h:12 h light–dark cycle) with food and water available ad libitum. Rats weighing between 150 and 180 g (approximately 5–6 weeks old) were used to induce the renovascular 2K1C Goldblatt hypertension model (RVH), a well‐characterized and widely used model. Rats were anaesthetized with isoflurane (3%) throughout the surgery. Following an abdominal incision, the left kidney was exposed, and a 0.2 mm clip was placed over the left renal artery, partially occluding it. Sham rats were subjected to the same surgical procedure, although the artery was not occluded. Post‐operative care included proper management of associated pain (buprenorphine, 0.25 mg kg⁻¹, subcutaneous, as needed). Systolic blood pressure was measured at the beginning of the sixth week post‐surgery, using a tail‐cuff method. Mean systolic blood pressure values for sham and RVH were 139.9 ± 2.7 (n = 14) and 200.0 ± 3.1 mmHg (n = 16), respectively (P < 0.0001, Student's unpaired t test). All rats were used for experiments during the sixth to seventh week post‐surgery.

Hypothalamic slices

Rats were anaesthetized with pentobarbital (50 mg kg⁻¹ i.p.), quickly decapitated and brains dissected out. Coronal slices were cut (250 μm thick) utilizing a vibroslicer (Leica VT1200s, Leica Microsystems (Buffalo Grove, IL, USA)) as previously described (Fleming et al. 2011). An oxygenated ice cold artificial cerebrospinal fluid (ACSF) was used during slicing (containing in mm): 119 NaCl, 2.5 KCl, 1 MgSO₄, 26 NaHCO₃, 1.25 NaH₂PO₄, 20 d‐glucose, 0.4 ascorbic acid, 2.0 CaCl₂ and 2.0 pyruvic acid; pH 7.4; 295–305 mosmol l⁻¹. After sectioning, slices were placed in a holding chamber containing ACSF and kept at room temperature (22°C) until used.

Electrophysiological recordings

Slices were placed in a submersion‐style recording chamber, and bathed with solutions (3.0 ml min⁻¹) that were bubbled continuously with a gas mix of 95% O₂–5% CO₂, and maintained at near physiological temperature (32°C). Thin‐walled (1.5 mm o.d., 1.17 mm i.d.) borosilicate glass (G150TF‐3, Warner Instruments, Sarasota, FL, USA) was used to pull patch pipettes (3–5 MΩ) on a horizontal Flaming/Brown micropipette puller (P‐97, Sutter Instruments, Novato, CA, USA). Whole‐cell patch‐clamp recordings from SON neurons were visually made using differential interference contrast (DIC) videomicroscopy as previously described (Fleming et al. 2011). Recordings were obtained with a Multiclamp 700A amplifier (Axon Instruments, Union City, CA, USA). The voltage output was digitized at 16‐bit resolution, 10 kHz (Digidata 1440, Axon Instruments), and saved on a computer to be analysed offline using pCLAMP10 software (Axon Instruments). Mean series resistance was 12.5 ± 0.4 MΩ and experiments were discarded in cases in which the series resistance was unstable or changed >20%. The internal solution contained (in mm): 140 potassium gluconate, 0.2 EGTA, 10 HEPES, 10 KCl, 0.9 MgCl₂, 4 MgATP, 0.3 NaGTP and 20 phosphocreatine (Na⁺); pH was adjusted to 7.2–7.3 with 1 mm KOH and the osmolarity was 280–290 mosmol l⁻¹. The ACSF contained (in mm): 85.09 NaCl, 30 TEA, 2.5 KCl, 26 NaHCO₃, 1.25 NaH₂PO₄, 20 d‐glucose, 0.4 ascorbic acid, 2.0 CaCl₂ and 2.0 pyruvic acid. Tetrodotoxin (0.5 μm) was added to prevent generation of action potentials. All recordings were obtained in low Mg²⁺ (10 μm MgSO₄) and in the presence of glycine (10 μm) in order to facilitate measurements of NMDA currents (I _NMDA), as we previously reported (Fleming et al. 2011). I _NMDA was activated by focal, transient application of NMDA via a picospritzer. I _NMDA current densities were determined by dividing the current amplitude by the cell capacitance, obtained by integrating the area under the transient capacitive phase of a 5 mV depolarizing step pulse, in voltage clamp mode.

Confocal Ca²⁺ imaging

SON neurons were loaded through the patch pipette with fluo‐5F pentapotassium salt (100 μm; Thermo Fisher Scientific, Waltham, MA, USA). Once the whole‐cell configuration was established, the dye was allowed to dialyse into the cell for at least 15 min before the initiation of the recordings. Calcium imaging was conducted using the Yokogawa real time live cell laser confocal system (CSU‐10) combined with a highly sensitive EMCCD camera (iXon+ 885, Andor Technology, South Windsor, CT, USA) (Son et al. 2013). Fluorescence images were obtained using a diode‐pumped solid‐state laser (Melles Griot, Carlsbad, CA, USA), and fluorescence emission was collected at >495 nm. Images were acquired at a rate of 4 frames s⁻¹. The fractional fluorescence (F/F ₀) was determined by dividing the fluorescence intensity (F) within a region of interest (ROI; 10 × 10 pixels) by a baseline fluorescence value (F ₀) determined from 30 images before brief application of NMDA (a period showing no change in intracellular calcium levels), as previously described (Son et al. 2013). The ROI was positioned within the cell soma, and care was taken to avoid the cell nucleus. Data were analysed using Andor IQ software (Andor Technology) and Image J (NIH).

Statistical analysis

All values are expressed as means ± SEM. Drug effects in sham and RVH rats were assessed using two‐way analysis of variance repeated measures (ANOVA‐RM), as indicated. Where the F ratio was significant, post hoc comparisons were completed using the Bonferroni post hoc test. When needed, and as indicated, Student's unpaired t test was used to compare basal differences between sham and RVH rats. Pearson's correlation test was used to determine if correlations existed between two parameters. Differences were considered statistically significant at P < 0.05 and n refers to the number of cells. Statistical significance was tested at the 95% (P < 0.05) confidence level. All statistical analyses were conducted using GraphPad Prism (GraphPad Software, San Diego, CA, USA).

Results

NMDA‐evoked currents do not differ between sham and RVH rats

To analyse and compare the properties of NMDA currents in MNCs from sham and RVH rats, patched MNCs were held at various membrane potentials and subjected to repetitive focal applications of NMDA (100 μm, 500 ms duration). Mean cell capacitance (sham: 38.5 ± 5.8 pF, RVH: 37.7 ± 1.0 pF) and input resistance (sham: 796.5 ± 58.3 MΩ, RVH: 868.3 ± 32.9 MΩ)  did not differ between experimental groups. As shown in the representative examples of Fig. 1 A NMDA evoked fast inward currents from holding potentials of −50 mV to 0 mV, reversing to outward currents at positive membrane potentials. Figure 1 B shows plots of the mean current amplitude and density, respectively, as a function of the holding potential. Results from a two‐way ANOVA indicated an overall small though significantly diminished I _NMDA peak amplitude in RVH compared to sham rats (F = 1.6, P < 0.05, n = 6 in each group). However, post hoc analysis indicated that no individual differences were observed for any of the voltages measured. When data were expressed as current density (i.e. normalized by cell capacitance), similar results were observed (F = 1.6, P < 0.05). Conversely, no differences in I _NMDA area were observed between groups (F = 0.0001, P = 0.9, not shown). As we recently reported, I _NMDA reversal potential was not different between MNCs in sham and RVH rats (Zhang et al. 2017).

Figure 1. NMDAR activation evokes similar currents in MNCs from sham and RVH rats.

A, representative examples of evoked NMDAR‐mediated currents (I _NMDA) in a MNC from a sham (Aa) and an RVH (Ab) rat, following a puff of NMDA (picospritzer, 100 μm, 4 p.s.i., 500 ms), while holding the neurons at varying holding potentials (from −50 to +30 mV in 10 mV increments). B, mean plots of I _NMDA peak amplitude (Ba) and current density (Bb) as a function of the holding Vm in sham and RVH rats (n = 6 each).

The time course of NMDAR‐evoked changes in intracellular somatic Ca²⁺ levels is prolonged in RVH rats

In addition to evoking direct membrane depolarization, activation of NMDARs leads to an increase in intracellular levels of Ca²⁺. Thus, to compare the NMDAR‐evoked changes in intracellular Ca²⁺ (NMDAR‐ΔCa²⁺) between MNCs in sham and RVH rats, we obtained simultaneous patch‐clamp and confocal Ca²⁺ imaging recordings in MNCs loaded with fluo‐5 (Methods). In order to better detect and assess changes in intracellular Ca²⁺ dynamics, longer puffs of NMDA, but at a lower concentration, were used (20 μm, 5 s). As shown in Fig. 2, focal application of NMDA onto the somata of the recorded neurons evoked a robust increase in somatic [Ca²⁺]_i whose time course surpassed the duration of the evoked I _NMDA. We found in this set of studies, similar to what we reported in Fig. 1, that the peak amplitude, area and decay kinetics of the focally evoked I _NMDA was not different between MNCs in sham and RVH rats (n = 24 in each group) (Fig. 3 A–C).

Figure 2. Representative examples of NMDAR‐evoked increase in intracellular Ca²⁺ levels in sham and RVH rats.

A, simultaneous I _NMDA (voltage clamp, lower traces) and intracellular Ca²⁺ (upper traces) measurements in response to a focal brief application of NMDA (50 μm, 4 p.s.i., 5 s) in a sham (Aa) and RVH (Ab) rat. B, representative confocal Ca²⁺ images (pseudocolour) of the patched neurons at the time points (1–3/1′–3′) indicated in the ΔCa²⁺ traces are shown below. Scale bar in Ba: 15 μm. Arrows in A point to the time of the NMDA puffs. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 3. NMDAR‐evoked increase in intracellular Ca²⁺ is prolonged in MNCs from RVH rats.

Summary of the mean I _NMDA peak amplitude (A), area (B) and decay time constant (τ) (C) in MNCs from sham and RVH rats. Summary of the mean ΔCa²⁺ peak (D), decay time constant (τ) (E), % residual Ca²⁺ (F) and area (G) in MNCs from sham and RVH rats. H, plot of ΔCa²⁺ area as a function of I _NMDA area in MNCs from sham and RVH rats. n = 24 in each group. ^* P < 0.05 and ^** P < 0.01 vs. Sham, unpaired t test.

Measurements of the NMDAR‐ΔCa²⁺ signal revealed no significant differences in the [Ca²⁺]_i signal peak amplitude between groups (Fig. 3 D). However, the decay time course of the NMDAR‐ΔCa²⁺ signal following the NMDA stimulus was notably slower in MNCs from RVH rats. Thus, the decay time constant (τ) in the latter was significantly longer compared to MNCs in sham rats (P < 0.01, Fig. 3 E). In many cases, as shown in the representative trace in Fig. 2 B, [Ca²⁺]_i did not return completely to the baseline level before NMDAR stimulation, at least within the time period of the recording. Thus, when this phenomenon was quantified as the percentage NMDAR‐ΔCa²⁺ residual at the end of the recording, relative to the [Ca²⁺]_i peak, a significantly larger residual Δ[Ca²⁺]_i was observed in MNCs from RVH compared to sham rats (P < 0.02, Fig. 3 F). The slower decay time course of the NMDAR‐ΔCa²⁺ signal in MNCs from RVH rats resulted consequently in a significantly larger NMDAR‐ΔCa²⁺ signal area compared to those in sham rats (P < 0.05, Fig. 3 G). We found no significant correlations between the magnitude of I _NMDA and the magnitude of the evoked [Ca²⁺]_i (R ² = 0.05 and 0.18 for sham and RVH, respectively (Fig. 3 H).

In a few instances, we observed a more complex response to the NMDA stimulation, consisting of an initial fast inward component, reflecting the rapid summation of excitatory postsynaptic currents (EPSCs). This response was then followed by a slower, and more persistent I _NMDA component. These two phases of the I _NMDA response were also clearly reflected in the evoked NMDAR‐ΔCa²⁺ signal, in which concomitant fast and slow components were also observed (Fig. 4). These responses were equally observed both in sham and RVH rats (n = 6/group).

Figure 4. Complex I _NMDA and ΔCa²⁺ responses following NMDAR activation in a subset of MNCs.

A, representative example of simultaneous I _NMDA (voltage clamp, lower trace) and intracellular Ca²⁺ (upper trace) measurements in response to a focal brief application of NMDA, in which an initial fast inward component, likely representing summation of fast EPSPs (arrow, see also expanded inset corresponding to the underlined part of the main trace) is followed by a slower, longer lasting response. These two clearly separate components in I _NMDA were also reflected as two separate peaks in the intracellular ΔCa²⁺ response. B, representative confocal Ca²⁺ images (pseudocolour) of the patched neuron at the time points (1 and 2) indicated in the ΔCa²⁺ traces are shown. Scale bar = 10 μm. The arrow in A points to the time of the NMDA puff. [Color figure can be viewed at wileyonlinelibrary.com]

Changes in NMDAR‐evoked increases in dendritic intracellular Ca²⁺ levels in RVH rats

In a limited number of cases (n = 5 and 7 in sham and RVH rats, respectively), intracellular loading with fluo5‐F was efficient enough to load and visualize neuronal dendrites, allowing us to monitor also dendritic changes in NMDAR‐ΔCa²⁺ signal. A representative image is shown in Fig. 5. For these recordings, we quantified changes at the most distal segment of the visible dendrite. This length ranged from 12.6 to 160.8 μm. Similar to what we observed for the soma, a significantly prolonged decay time course and a larger NMDAR‐ΔCa²⁺ area were observed (P < 0.02 for both τ and area) (Fig. 5 C). As shown in Fig. 5 D, no significant correlations between the dendritic length and ΔCa²⁺ peak (R ² = 0.25 sham and 0.15 RVH), area (R ² = 0.50 sham and 0.01 RVH) or decay τ (R ² = 0.001 sham and 0.13 RVH) were observed (P > 0.2 in all cases).

Figure 5. NMDAR‐evoked increase in dendritic intracellular Ca²⁺ is prolonged in MNCs from RVH rats.

Aa and b, representative confocal Ca²⁺ images (pseudocolour) of a fluo‐5‐loaded MNC before (Aa) and after (Ab) a puff of NMDA (50 μm, 4 p.s.i., 5 s). Ac, somatic and dendritic ΔCa²⁺ measurements (F/F ₀) obtained from the small squared areas indicated in Aa and Ab, respectively. Ba, an expanded dendritic segment (black rectangular area in Ab) obtained at the peak of the ΔCa²⁺ response is shown. Note the three dendritic varicosities (1–3), which showed the largest ΔCa²⁺ response, as shown quantitatively in Bb, which represents a line scan plot (F/F ₀ as a function of distance) along that dendritic segment. Ca–c, summary of the mean dendritic ΔCa²⁺ peak (Ca), area (Cb) and decay time constant (τ) (Cc) for MNCs in sham (n = 5) and RVH (n = 7) rats. Da–c, plots of ΔCa²⁺ peak (Da), area (Db) and decay time constant (τ) (Dc) as a function of the dendritic length from soma. ^* P < 0.02 compared to Sham, unpaired t test. [Color figure can be viewed at wileyonlinelibrary.com]

Interestingly, line scans obtained from a subset of dendrites in these recordings indicated that the most predominant dendritic ΔCa²⁺ signals following NMDAR activation occurred within defined dendritic “hot spots”, which likely represent dendritic varicosities (see representative example in Fig. 5 B).

A blunted endoplasmic reticulum buffering capacity contributes to the prolonged NMDAR‐ΔCa²⁺ in MNCs from RVH rats

A major factor influencing the shape and time course of intracellular Ca²⁺ in MNCs is buffering by the endoplasmic reticulum (ER) via the sarco/endoplasmic reticulum Ca²⁺‐ATPase (SERCA) pump (Toescu & Verkhratsky, 1998). Thus to determine whether the prolonged NMDAR‐ΔCa²⁺ in MNCs from RVH rats was due to a diminished ER Ca²⁺ buffering capacity, we repeated a set of experiments in slices preincubated with thapsigargin (TG) (3 μm, 45 min), an endoplasmic reticulum SERCA pump blocker (Lytton et al. 1991) that leads to depletion of ER Ca²⁺ stores.

To first test the efficacy of TG in our experimental conditions, MNCs from slices incubated in the presence or absence of TG were exposed to a focal application of caffeine (50 μm, 5 s), known to mobilize Ca²⁺ from ER (Komori et al. 2010). As shown in Fig. 6, the magnitude of the caffeine‐evoked ΔCa²⁺ signal was significantly blunted in slices preincubated in TG (P < 0.0001, n = 7 and 10 in control and TG, respectively), supporting that TG significantly blocked SERCA pump activity and ER Ca²⁺ uptake.

Figure 6. Thapsigargin (TG) blunts caffeine‐induced increases in intracellular Ca²⁺ levels.

Aa and b, representative confocal Ca²⁺ images (pseudocolour) of a fluo‐5‐loaded MNC before (Aa) and after (Ab) a puff of caffeine (50 μm, 5 s) in control ACSF. Ac, somatic ΔCa²⁺ measurements (F/F ₀) in response to caffeine obtained from the same MNC shown in Aa and b. Ba and b, representative confocal Ca²⁺ images (pseudocolour) of a different fluo‐5‐loaded MNC before (Ba) and after (Bb) a puff of caffeine (50 μm, 5 s) in a slice preincubated in TG (3 μm, 45 min). Bc, somatic ΔCa²⁺ measurements (F/F ₀) in response to caffeine obtained from the same MNC shown in Ba and b. C, summary data of mean ΔCa²⁺ peak amplitude in response to caffeine in slices incubated in normal ACSF (n = 7) or in TG (n = 10). ^*** P < 0.0001 vs. ACSF unpaired t test. Scale bar in Bb = 20 μm. [Color figure can be viewed at wileyonlinelibrary.com]

In separate sets of studies, we then repeated focal applications of NMDA to MNCs from sham (n = 19) and RVH (n = 26) rats in slices preincubated in TG as above. Representative examples of these recordings are shown in Fig. 7 A and B. Importantly, we found that in slices preincubated in TG, the decay time course of the NMDAR‐ΔCa²⁺ in sham rats was slowed down, to the same level as we previously found in RVH rats. Moreover, no further increases in the decay τ were observed in RVH rats in the TG group (Fig. 7 Ca). Similar results were observed when the NMDAR‐ΔCa²⁺ magnitude (area) was compared among groups (Fig. 7 Cb). Thus, the prolonged and enhanced NMDAR‐ΔCa²⁺ observed in RVH rats in control conditions was absent in slices preincubated in TG. Unfortunately, we were unable to obtain in this set of studies sufficient MNCs with dendritic labelling, which precluded us from obtaining a similar analysis to that shown in Fig. 5.

Figure 7. Thapsigargin (TG) blunts caffeine‐induced increases in intracellular Ca²⁺ levels.

Aa and b, representative confocal Ca²⁺ images (pseudocolour) of a fluo‐5‐loaded MNC in a sham rat before (Aa) and after (Ab) a puff of NMDA (50 μm, 5 s) in slices preincubated in TG. Ac, somatic ΔCa²⁺ measurements (F/F ₀) in response to NMDA obtained from the same MNC shown in Aa and b. Ba and b, representative confocal Ca²⁺ images (pseudocolour) of a fluo‐5‐loaded MNC in an RVH rat before (Aa) and after (Ab) a puff of NMDA (50 μm, 5 s) in slices preincubated in TG. Bc, somatic ΔCa²⁺ measurements (F/F ₀) in response to NMDA obtained from the same MNC shown in Ba and b. Ca and b, combined summary data of mean ΔCa²⁺ decay time constant (τ) and area, respectively, in MNCs from sham and RVH rats in control ACSF (n = 24 and 24, respectively, for sham and RVH) and in slices preincubated in TG (n = 19 and 26, respectively, for sham and RVH, red columns). Please note that the data for control ACSF is the same as shown in Fig. 3; it was included here for better comparison with data obtained in TG. ^* P < 0.05, ^** P < 0.01 and ^*** P < 0.001 vs. Sham, Bonferroni's post hoc test. Scale bar in Ab = 20 μm. [Color figure can be viewed at wileyonlinelibrary.com]

Discussion

Using a combination of patch‐clamp electrophysiology with simultaneous fast confocal Ca²⁺ imaging in a renovascular hypertensive rat model (RVH), we aimed in this study to determine whether NMDAR‐evoked changes in intracellular Ca²⁺ dynamics (NMDAR‐ΔCa²⁺) in MNCs was altered during hypertension, and if so, to identify potential underlying mechanisms. Our main findings indicate that (1) despite a similar magnitude of NMDA receptor‐mediated currents (I _NMDA), an altered spatiotemporal profile in the ΔCa²⁺ response was evoked in MNCs of hypertensive rats; (2) the NMDAR‐ΔCa²⁺ response was significantly slower and larger over time, resulting in an overall exacerbated ΔCa²⁺ response per unit of I _NMDA; (3) the prolonged and larger NMDA‐ΔCa²⁺ responses in MNCs of RVH rats affected both somatic and dendritic compartments; (4) inhibition of the ER SERCA pump activity with thapsigargin prolonged NMDAR‐ΔCa²⁺ responses in MNCs of sham rats, but this effect was occluded in RVH rats, equalizing the responses observed in sham and RVH rats in this condition. Taken together, these results support an exacerbated NMDAR‐ΔCa²⁺ signalling in MNCs of hypertensive rats, and suggest that a blunted ER Ca²⁺ uptake capacity is a contributing underlying mechanism to this effect.

Exacerbated NMDAR‐ΔCa²⁺ signalling in RVH rats: role of altered ER Ca²⁺ buffering capacity

Several studies from our laboratory and others support an exacerbated glutamatergic function in the SON/PVN of hypertensive rats (Li & Pan, 2007; Zheng et al. 2011; Gabor & Leenen, 2012; Glass et al. 2015). However, the precise underlying mechanisms and pathways ultimately leading to increased neuronal activity and increased neurohumoral outflow from the SON/PVN during this condition remain to be fully elucidated. Here, we focused on addressing the contribution of changes in the spatiotemporal pattern of intracellular Ca²⁺ dynamics, a critical process in determining the overall efficacy and downstream actions on membrane excitability that follows the activation of glutamate NMDARs.

One of the key findings in this study is that despite a similar I _NMDA magnitude between MNCs from sham and RVH rats, a more prolonged and larger NMDAR‐ΔCa²⁺ was observed in the latter. The spatiotemporal pattern and magnitude of the changes in intracellular Ca²⁺ levels that follow the activation of NMDARs is influenced by a number of different factors, including the expression density of NMDA receptors, the single‐channel NMDAR Ca²⁺ permeability, as well as intracellular Ca²⁺ buffering mechanisms. Since Ca²⁺ influx is a major component mediating I _NMDA (McBain & Mayer, 1994), and since the magnitude of I _NMDA was not different between sham and RVH rats, it is reasonable to argue that an increase in NMDA receptor numbers and/or in Ca²⁺ permeability were not factors contributing to the larger NMDAR‐ΔCa²⁺ observed in MNCs from RVH rats.

Following an initial rise in intracellular Ca²⁺ levels, Ca²⁺ buffering and cytosolic clearance mechanisms gradually decrease cytosolic free Ca²⁺ levels, resulting in a slowly decaying Ca²⁺ time course, Thus, a diminished efficacy of Ca²⁺ buffering mechanisms is expected to mostly affect the decay phase of the Ca²⁺ transient. In MNCs, the endoplasmic reticulum Ca²⁺ transport ATPase (ER‐SERCA) has been shown to efficiently shape somatic Ca²⁺ transients (Kim et al. 2003; Komori et al. 2010; Dayanithi et al. 2012). Thus, to determine whether a blunted ER‐SERCA buffering mechanism contributed to the prolonged NMDA‐ΔCa²⁺ in RVH rats, we compared the effects of thapsigargin (TG), a SERCA blocker, on the NMDAR‐ΔCa²⁺ signalling both in sham and RVH rats. We found that in MNCs from sham rats, TG significantly prolonged the decay time course and increased the overall magnitude of the NMDAR‐ΔCa²⁺ signal to a similar extent to that observed under control conditions in MNCs of RVH rats. Conversely, TG failed to affect the NMDAR‐ΔCa²⁺ signal in RVH rats. Taken together, these results are in line with the notion that (1) ER‐SERCA efficiently restricts the magnitude and time course of NMDAR‐ΔCa²⁺ signalling in MNCs, and (2) that a blunted ER‐SERCA buffering capacity contributes to the prolonged NMDAR‐ΔCa²⁺ signal observed in MNCs in RVH rats. Other alternative mechanisms, however, including changes in NMDAR desensitization properties, known to affect I _NMDA decay kinetic properties (Jonas & Spruston, 1994) could not be completely ruled out.

It is important to acknowledge that the ER could also act as a source of intracellular free Ca²⁺ levels by releasing stored Ca²⁺, in most instances, in a Ca²⁺‐dependent manner (i.e. Ca²⁺‐induced Ca²⁺ release) (Hongpaisan et al. 2001). In fact, NMDAR activation has been shown to evoke ER‐Ca²⁺ release in hippocampal neurons, contributing to the overall NMDAR‐ΔCa²⁺ waveform (Emptage et al. 1999). In our study, however, TG treatment never resulted in a reduction of the NMDAR‐ΔCa²⁺ signal. Taken together, our study indicates that in MNCs, the ER takes up (but does not release) Ca²⁺ in response to NMDAR activation.

The precise mechanism ultimately leading to altered ER Ca²⁺ buffering and enhanced NMDAR‐ΔCa²⁺ signalling in RVH rats is at present unknown. ER stress, a process that involves the accumulation of unfolded or misfolded proteins in the ER lumen, has been proposed to be a critical mechanism in the development and maintenance of neurogenic hypertension (Young et al. 2012; Chao et al. 2013). ER stress could result from various pathological processes, including angiotensin II‐mediated oxidative stress (Chao et al. 2013), or subsequent to a cytoplasmic Ca²⁺ overload (Richter et al. 2016). In fact, sustained NMDAR activation in hippocampal neurons was recently shown to contribute to ER stress via Ca²⁺ overload (Dong et al. 2017). Conversely, once established, ER stress affects ER Ca²⁺ homeostasis, furthering in turn cytosolic Ca²⁺ accumulation (Richter et al. 2016). Nevertheless, whether the enhanced tonic NMDAR activation that occurs in MNCs during hypertension (Zhang et al. 2017) is a critical factor itself contributing to ER stress and disrupted intracellular Ca²⁺ homeostasis in this condition, remains to be determined.

As reported, we observed in some instances a multimodal response to NMDA consisting of an initial fast inward component that reflected the rapid summation of excitatory postsynaptic currents (EPSCs) followed by the most typical slower, and more persistent I _NMDA component. While this phenomenon was not further explored in this work, the initial fast component could reflect activation of presynaptic NMDARs resulting in enhanced endogenous neurotransmitter release efficacy, as reported elsewhere (Bouvier et al. 2015).

In relation to this, one limitation of our study is that a puff of NMDA, due to the relatively slow and diffuse spatiotemporal profile of agonist delivery, results most likely in the activation of extrasynaptic over synaptic NMDARs, largely because the latter are contained within a spatially restricted structure, and desensitize much more rapidly than eNMDARs, which can remain activated in a sustained manner in the presence of the agonist (Sah et al. 1989; Okamoto et al. 2009). We recently showed that eNMDARs carry the majority of the charge transfer mediated by glutamate excitatory actions in MNCs, strongly stimulating activity in these neurons (Fleming et al. 2011). Moreover, we recently showed that eNMDARs (but not synaptic NMDARs) are functionally coupled, in a Ca²⁺‐dependent manner, to the A‐type K⁺ channel‐mediated current I _A (Naskar & Stern, 2014) and that an augmented eNMDAR–I _A coupling is a critical mechanism contributing to exacerbated MNC activity in RVH rats (Zhang et al. 2017). Thus, the fact that we activated eNMDARs in this study is functionally relevant in the context of these previous studies. However, it remains unknown whether glutamate changes in intracellular Ca²⁺ following activation of synaptic NMDARs would be similarly affected in RVH rats.

Another limitation of our study is that we did not distinguish between vasopressin (VP) and oxytocin (OT) MNC phenotypes. Nevertheless, the NMDAR‐ΔCa²⁺ signals in sham and RVH rats were relatively consistent across all MNCs tested in each group, suggesting that both cell types were likely affected in hypertensive rats. Future studies will be needed, however, to determine if NMDAR‐ΔCa²⁺ signals in VP and OT MNCs are differentially affected in RVH rats.

Functional consequences of altered NMDAR‐ΔCa²⁺ signalling in RVH rats

Prolonged and larger NMDAR‐ΔCa²⁺ signalling could have numerous downstream consequences on MNC excitability in RVH rats. Critical ion channels and neurotransmitter pathways that influence MNC activity and neurosecretory outflow are modulated in a Ca²⁺‐dependent manner. Thus, a rise in intracellular Ca²⁺ levels following NMDAR activation could lead to activation of inhibitory channels, such as the small conductance Ca²⁺‐dependent K⁺ channels (SK), or to excitatory channels such as Ca²⁺‐activated cation (CAN) channels, including TRPM5/6 (Teruyama et al. 2011), which would act in a negative or positive feedback, respectively, to the NMDAR‐mediated excitatory effect. We recently showed that NMDARs in MNCs are functionally coupled to A‐type K⁺ channels (Naskar & Stern, 2014), which efficiently affect MNC firing by regulating spike onset and interspike interval during repetitive firing (Bourque, 1988; Luther & Tasker, 2000). Specifically, we showed that NMDAR activation leads to a Ca²⁺‐dependent inhibition of A‐type K⁺ channels, an effect that potentiates the excitatory effect of NMDAR activation (Naskar & Stern, 2014). Moreover, we also reported that an augmented NMDAR tone contributed to blunted A‐type K⁺ channel activity, and thus firing discharge in MNCs in RVH rats (Zhang et al. 2017). Thus, the results from the present work support the notion that a blunted ER Ca²⁺ buffering capacity could be a contributing mechanism leading to the enhanced NMDAR‐I _A coupling and exacerbated firing activity of MNCs during hypertension. Clearly, future studies are warranted to further assess how the various downstream Ca²⁺‐dependent signalling mechanisms are affected in MNCs during hypertension, as a consequence of the enhanced NMDAR‐ΔCa²⁺ signalling in this condition.

We found that the increased NMDAR‐ΔCa²⁺ signal observed in MNCs from RVH rats was not only restricted to the soma, but also involved dendritic compartments. In MNCs, this phenomenon is expected to have important physiological consequences. VP and OT are not only released from neurohypophysial terminals into the circulation, but also intranuclearly from the dendrites of MNCs (Ludwig & Leng, 2006). Dendritic release of these neuropeptides is activity‐ and Ca²⁺‐dependent (Ludwig & Leng, 2006), and activation of NMDARs, via the subsequent increase in intracellular Ca²⁺, serves as a powerful mechanism to trigger dendritic release (de Kock et al. 2004; Son et al. 2013). Intranuclear release of VP and OT acts as a powerful autoregulatory mechanism to regulate MNC firing activity and to optimize systemic release (Ludwig & Leng, 1997; Gouzenes et al. 1998). Moreover, we recently showed that dendritic release of VP, via diffusion in the extracellular space, stimulates the activity of neighbouring presympathetic neurons, increasing sympathoexcitatory outflow from the PVN (Son et al. 2013). Thus, the enhanced NMDAR‐ΔCa²⁺ in RVH rats may lead to an augmented dendritic VP release, contributing in turn to enhanced neurohumoral outflow during hypertension.

In summary, results from the present study show that NMDAR‐ΔCa²⁺ signalling in MNCs is exacerbated in MNCs from RVH rats, and that a blunted ER Ca²⁺ buffering capacity contributes as an underlying mechanism. Future studies are warranted to further assess the precise mechanisms leading to altered ER function in RVH rats as well as the overall contribution of this mechanism to neurohumoral activation in neurogenic hypertension.

Additional information

Competing interests

None declared.

Author contributions

M.Z. carried out the experiments, analysed, interpreted data and contributed to manuscript writing. J.E.S. advised on the experiments and the analysis, and wrote the manuscript. All authors have approved the final version of the manuscript and agree to be accountable for all aspects of the work. All persons designated as authors qualify for authorship and all those who qualify for authorship are listed.

Funding

This work was supported by an NIH National Heart, Lung, and Blood Institute grant (R01 HL112225) to J.E.S.