RyR-mediated Ca²⁺ release elicited by neuronal activity induces nuclear Ca²⁺ signals, CREB phosphorylation, and Npas4/RyR2 expression

Significance

The expression of genes involved in hippocampal synaptic plasticity, learning, and memory requires that Ca²⁺ signals generated in spines, dendrites, or the soma by neuronal stimulation reach the nucleus. Here, we report that stimulation of hippocampal neurons induces Ca²⁺ release mediated by RyR2 channels, which contributes to nuclear Ca²⁺ signal generation. Suppression of RyR-mediated Ca²⁺ release inhibited the activity-induced phosphorylation of the nuclear transcriptional regulator CREB and the expression of the Npas4 transcription factor and RyR2, which play crucial roles in hippocampal memory processes. We propose that RyR-mediated Ca²⁺ release induced by neuronal stimulation, by promoting the sequential generation of nuclear Ca²⁺ signals, CREB phosphorylation, and Npas4/RyR2 up-regulation, plays a significant role in hippocampal synaptic plasticity and spatial memory processes.

Abstract

The expression of several hippocampal genes implicated in learning and memory processes requires that Ca²⁺ signals generated in dendritic spines, dendrites, or the soma in response to neuronal stimulation reach the nucleus. The diffusion of Ca²⁺ in the cytoplasm is highly restricted, so neurons must use other mechanisms to propagate Ca²⁺ signals to the nucleus. Here, we present evidence showing that Ca²⁺ release mediated by the ryanodine receptor (RyR) channel type-2 isoform (RyR2) contributes to the generation of nuclear Ca²⁺ signals induced by gabazine (GBZ) addition, glutamate uncaging in the dendrites, or high-frequency field stimulation of primary hippocampal neurons. Additionally, GBZ treatment significantly increased cyclic adenosine monophosphate response element binding protein (CREB) phosphorylation—a key event in synaptic plasticity and hippocampal memory—and enhanced the expression of Neuronal Per Arnt Sim domain protein 4 (Npas4) and RyR2, two central regulators of these processes. Suppression of RyR-mediated Ca²⁺ release with ryanodine significantly reduced the increase in CREB phosphorylation and the enhanced Npas4 and RyR2 expression induced by GBZ. We propose that RyR-mediated Ca²⁺ release induced by neuronal activity, through its contribution to the sequential generation of nuclear Ca²⁺ signals, CREB phosphorylation, Npas4, and RyR2 up-regulation, plays a central role in hippocampal synaptic plasticity and memory processes.

Calcium-dependent changes in gene expression induced by neuronal activity participate in a wide range of processes and behaviors including persistent synaptic plasticity, dendritic structural changes, and memory consolidation (1, 2). In response to neuronal stimulation, Ca²⁺ influx via voltage-gated Ca²⁺ channels (VGCCs) or N-methyl-d-aspartate receptors (NMDARs) has a central role in neuronal excitation–transcription coupling in different neuronal types (1, 3). In hippocampal neurons, stimulation-induced postsynaptic Ca²⁺ signals propagate to the cell nucleus, where these signals exert a key function in the regulation of gene expression and play a crucial role in activity-dependent transcription (2–4). Activity-generated nuclear Ca²⁺ signals—defined as transient increments in the free Ca²⁺ concentration inside the nucleus—promote neuronal gene expression by stimulating diverse signaling cascades, including de novo induction of DNA methylation that conditions global chromatin remodeling and gene transcription (3, 5, 6). Thus, neuronal stimulation induces nuclear Ca²⁺ signals that translate synaptic stimuli into functional and structural neuronal changes.

The activation by nuclear Ca²⁺ signals of protein kinases such as CaMKII and CaMKIV induces phosphorylation/activation and/or changes in the localization of diverse transcriptional regulators (2). Following neuronal activation, these and other kinases target several transcription factors that include, among others, the cyclic adenosine monophosphate response element binding protein (CREB) and the Neuronal Per Arnt Sim domain protein 4 (Npas4) (7, 8). The nuclear-resident transcription factor CREB modulates many functions in the central nervous system, including synaptic plasticity, memory formation, and neurogenesis (9, 10). Likewise, nuclear Ca²⁺ signals specifically activate the neuronal transcription factor Npas4, which participates in diverse neuronal functions; it coordinates the redistribution of hippocampal inhibitory synapses from the apical dendrites to the cell body (11), it links neuronal activity to memory processes and neuroprotection (2, 6, 11–13), and it partakes in reward-related gene expression and behavior (14).

Synaptic NMDAR activation generates Ca²⁺ signals that induce CREB phosphorylation via the joint action of nuclear CaMKIV, which mediates rapid CREB phosphorylation, and the Ras-ERK1/2 pathway that promotes the slower but longer-lasting CREB phosphorylation required for gene expression changes (15, 16). Moreover, CREB-mediated transcription requires persistent CREB phosphorylation plus CaMKIV-mediated phosphorylation of CREB binding protein (CBP), a reaction that requires nuclear Ca²⁺ signals (17). In addition, VGCCs generate Ca²⁺ signals in neuronal cells (18), which contribute to activity-dependent regulation of gene expression (7, 18–20). Stimulation of Ca²⁺ entry through VGCCs or NMDARs in hippocampal neurons generates Ca²⁺ signals, which propagate to the nucleus, where they regulate diverse gene transcription pathways (18, 19).

Nuclear Ca²⁺–dependent gene transcription following synaptic activation requires that the local Ca²⁺ signals generated at postsynaptic spines, dendrites, or the soma reach the nucleus. Diffusion of Ca²⁺ in the cytoplasm is highly restricted (20, 21), so hippocampal neurons must use other mechanisms to propagate activity-generated local Ca²⁺ signals to the nucleus. At present, the cellular mechanisms underlying this propagation remain unresolved, albeit several mechanisms have been proposed (1, 2, 22). Nuclear Ca²⁺ signals may arise from the diffusion and/or active transport and subsequent nuclear translocation of Ca²⁺-regulated proteins from their activation sites into the nucleus (23–25). Propagation of inositol 1,4,5-trisphosphate (IP₃) receptor-dependent waves of internal Ca²⁺ release to the somato-nuclear compartment represents another possible mechanism (26, 27). A third proposal considers that Ca²⁺ waves generated via Ca²⁺-induced Ca²⁺ release (CICR) mediate the propagation of Ca²⁺ signals from their sites of origin to the nucleus (28).

Relatively few studies have addressed the participation of CICR in the generation of nuclear Ca²⁺ signals in neuronal cells and the resulting gene expression changes. Calcium release from intracellular stores mediates CREB-dependent gene expression in hippocampal neurons stimulated with NMDA (29). Likewise, action potentials generated at the soma of spinal cord neurons generate cytoplasmic and nuclear Ca²⁺ signals, which are mediated by the activation of VGCCs and the subsequent amplification by Ca²⁺ release from ryanodine-sensitive intracellular stores (30). Conversely, studies in rat superior cervical ganglion neurons have shown that RyR channel inhibition does not affect signaling to CREB induced by activation of VGCCs (31). Moreover, suppression of RyR channel activity in hippocampal neurons only partially reduces the cytoplasmic Ca²⁺ increases elicited by glutamate uncaging in dendrites and soma and does not prevent Ca²⁺ signal propagation along the dendrite to the soma (32). In contrast, other studies have reported that RyR-mediated Ca²⁺ release plays a significant role in the generation of nuclear Ca²⁺ signals elicited by NMDAR’s activation with the GABA_A receptor blocker bicuculline, which promote CREB-mediated gene transcription (15, 29). Furthermore, in hippocampal slices, glutamate uncaging in neuronal dendrites triggers NMDAR activation that promotes ryanodine-sensitive intracellular Ca²⁺ release, which contributes to synapse development in CA1 pyramidal neurons in young but not in older rats; this coupling is not affected by pharmacological inhibition of VGCCs (33).

The contribution of RyR2 channels to nuclear Ca²⁺ signal generation has not been reported, even though the RyR2 isoform plays key roles in hippocampal structural plasticity and spatial memory processes (34). Moreover, RyR2 loss impairs neuronal activity-dependent remodeling of dendritic spines and triggers compensatory neuronal hyperexcitability (35). We have reported previously that the RyR2 protein content is higher than that of RyR3 and IP₃R1 both in rat hippocampal primary cultures and in the whole rat hippocampus (36). Here, we present results showing that stimulation of primary hippocampal neurons with three different protocols—gabazine (GBZ) addition, glutamate uncaging at the dendrites, or high-frequency field stimulation (HFFS)—engaged Ca²⁺ release mediated by RyR2 channels in the generation of nuclear Ca²⁺ signals. We also found that suppression of RyR-mediated Ca²⁺ release inhibited the CREB phosphorylation increase and the enhanced expression of Npas4 and RyR2 channels induced by GBZ. These results add to previous reports showing the relevance of nuclear Ca²⁺ signals, generated by local or global stimulation, in enhancing CREB phosphorylation (37) and the expression of genes required for synaptic plasticity and memory processes (38).

Materials and Methods

Plasmids and Antibodies.

A detailed list of the plasmids and antibodies used, and a detailed description of the transfection, immunofluorescence, qPCR protocols, and Western blot analysis used in this work, is provided in SI Appendix, Materials and Methods.

Primary Rat Hippocampal Cultures.

Sprague Dawley rats at 18 d gestation were used to prepare primary hippocampal cultures, as described (36). Briefly, hippocampal cells were recovered from the isolated embryonic hippocampi and were plated in Petri dishes previously treated with poly-l-lysine (0.1 mg/mL). Cultures were maintained next in serum-free Neurobasal Medium containing B27 Supplement, Glutamax TM, penicillin (20 U/mL), and streptomycin (20 μg/mL) at 37 °C under 5% CO₂. Cultures were used between 12 and 16 d in vitro. All experimental protocols used in this work complied with the “Guiding Principles for Research Involving Animals and Human Beings” of the American Physiological Society and were approved by the Bioethics Committee on Animal Research, Faculty of Medicine, Universidad de Chile.

Recording of Nuclear and Cytoplasmic Ca²⁺ Signals Induced by Synaptic Activity.

To detect nuclear or cytoplasmic Ca²⁺ signals, primary hippocampal cultures were transfected with the genetically encoded nuclear Ca²⁺ sensor GCaMP3.NLS or with the cytoplasmic Ca²⁺ sensor GCaMP5. As an additional strategy to detect neuronal Ca²⁺ signals, primary cultures were loaded for 15 min at 37 °C with the Ca²⁺ probe Fluo 4-AM, prepared in Tyrode’s extracellular medium. To assess the role of RyR2 channels, primary cultures were transfected with a short hairpin RNA molecule for RyR2 (shRyR2) or with the scrambled (shScr) sequence used in our previous work (34). The GCaMP3.NLS signals were partially filtered to eliminate bleed-through by the signal emitted by the RFP tag present in the shScr and shRyR2 sequences; this process reduced signal amplitude, so we compared the results with their corresponding controls. Transfected neurons were recorded in Tyrode’s solution (mM: 129 NaCl, 5 KCl, 25 Hepes, pH 7.3, 30 glucose, 2 CaCl₂, 1 MgCl₂). To suppress RyR-mediated Ca²⁺ release, cultures were preincubated for 60 min with 50 µM ryanodine; this condition suppresses RyR activity but preserves the endoplasmic reticulum (ER) Ca²⁺ content (34). In all cases, cells were placed in the microscopy stage of a wide-field Zeiss Cell Observer epifluorescence microscope (Zeiss), using as objectives Plan-Neofluar 20×/0.4 or Plan Apochromat, 40×/1.3 water, with light source, 470-nm Colibri 2 light-emitting diode (LED)-based module, and a digital camera, electron-multiplying charge-coupled device (EMCCD) Evolve 512 delta (Teledyne Photometrics). All settings were adjusted to minimize bleaching and maximize acquisition frequency. After recording a stable baseline, 10 μM GBZ was added to the culture; glutamate (10 μM) was added 15 to 20 min after GBZ addition; ionomycin (3 μM) was added at the end of the experiment to visualize the maximum fluorescence signals.

Glutamate Uncaging.

The uncaging of 4-methoxy-7-nitroindolinyl (MNI)-caged glutamate was performed at the microscope stage in pyramidal neurons [identified morphologically (39)]; to this aim, an ultraviolet (UV) laser beam (Micropoint, Andor) was directed close to the surface of the major dendrite. Neurons transfected with GCaMP3.NLS, GCaMP5, and/or shRyR2 or shScr were recorded 24 to 48 h after transfection. Uncaging was performed in Tyrode’s solution containing MNI-caged glutamate (2.5 mM). After baseline recording for 10 to 20 s, 5, 10, or 20 pulses of UV laser (365 nm) were applied at 15 Hz to a dendritic spot (∼3 to 4 µm in diameter). Fluorescence signals were recorded at 10 Hz in a spinning disk microscope (PerkinElmer/Zeiss) coupled to an Orca ER charge-coupled device (CCD) camera (Hamamatsu) controlled by commercial software (Volocity, PerkinElmer); a plan 40×/1.4 oil immersion objective was used to record the images.

Analysis of Ca²⁺ Signals.

Images were segmented by hand to generate binary masks suitable for morphological and functional analysis using the Fiji distribution of ImageJ software (40). All fluorescence signals were expressed as (F_max − F₀)/F₀, where F_max is the maximal fluorescence intensity and F₀ the intensity at the initial time (mean intensity of 20 to 50 frames recorded before the stimulus), both recorded in the region of stimulation. The maximum signal amplitude, F_max, was estimated by identifying the maximum intensity value induced by the stimulus.

Electrical Field Stimulation.

Neurons transfected with GCaMP3.NLS plus shRyR2 or shScr were recorded in Tyrode’s solution 24 to 48 h after transfection. Electrical field stimulation (30 V) was applied through platinum wires located in the microscope field of view using 1-ms suprathreshold voltage pulses delivered from high current capacity stimulators. Neurons were stimulated with trains of 50, 100, or 200 pulses at 10 or 20 Hz in the presence or absence of 50 µM ryanodine added to the Tyrode’s solution 60 min before applying the HFFS protocol. As previously described (41), the HFFS protocols used in this work do not produce neuronal death when evaluated 1, 6, and 24 h after the stimulus.

Results

Activity-Generated Cytoplasmic and Nuclear Ca²⁺ Signals.

Activity-induced cytoplasmic or nuclear Ca²⁺ signals were detected in primary hippocampal neurons transfected with genetically encoded cytoplasmic or nuclear Ca²⁺ sensors or in neurons loaded with Fluo-4. Neuronal activity was stimulated with GBZ, glutamate uncaging, or HFFS.

Stimulation of neuronal activity with GBZ.

Previous reports have shown that the GABA_A receptor inhibitors bicuculline or GBZ induce action potential bursting and elicit cytoplasmic and nuclear Ca²⁺ signals in primary hippocampal neurons (18, 42, 43). In agreement with these findings, we found that addition of 10 µM GBZ to primary hippocampal neurons promoted the emergence of synchronic oscillatory Ca²⁺ transients in the nucleus, as detected with the Ca²⁺ sensor GCaMP3.NLS (Fig. 1A). Due to its nuclear destination, the GCaMP3.NLS sensor exhibited at rest 50- to 100-fold higher fluorescence in the nucleus than in the cytoplasm (SI Appendix, Fig. S1 C and D). The addition of gabazine also generated cytoplasmic Ca²⁺ signals in both the dendrites and the soma, detected with the Ca²⁺ probe Fluo-4 (SI Appendix, Fig. S1 A and B and Movies S1 and S2) or with the GCaMP5 cytoplasmic Ca²⁺ sensor (Movie S3).

Fig. 1.

RyR-mediated Ca²⁺ release contributes to the generation of nuclear Ca²⁺ signals induced by GBZ. (A) Representative traces showing nuclear Ca²⁺ signals recorded in neurons transfected with GCaMP3.NLS before (baseline) and after the addition of 10 µM GBZ to primary hippocampal cultures. Synchronic oscillatory Ca²⁺ transients occur in the nucleus of hippocampal neurons expressing GCaMP3.NLS after GBZ addition. (B) Representative traces showing the effects of 10 µM GBZ addition on the nuclear Ca²⁺ signals exhibited by neurons expressing GCaMP3.NLS and incubated for 1 h with 50 µM ryanodine (Rya) to suppress RyR activity. (C) Average fluorescence values, expressed as (F_max − F₀)/F₀, were quantified in the nucleus following GBZ addition to control cultures or cultures treated for 1 h with 50 µM Rya. (D) Average values of the AUC, quantified in the nucleus following GBZ addition to control cultures or to cultures treated for 1 h with 50 µM Rya. (E) Representative traces showing the effects of 10 µM GBZ addition on nuclear Ca²⁺ signals recorded in neurons transfected with GCaMP3.NLS and cotransfected with a scrambled sequence, shScr. (F) Representative traces showing the effects of 10 µM GBZ addition on nuclear Ca²⁺ signals recorded in neurons transfected with GCaMP3.NLS and cotransfected with ShRyR2. (G) Average values of F_max, expressed as (F_max − F₀)/F₀, were quantified in the nucleus following GBZ addition to neurons expressing shScr. (H) Average values of the AUC quantified in the nucleus following GBZ addition to neurons expressing shRyR2. A period of 10 min after GBZ addition was taken as a time window to calculate AUC values. All individual representative traces correspond to paired conditions within the same culture; five independent cultures were analyzed. Data are expressed as mean ± SE; n = 5. Statistical analysis was performed with the unpaired Student’s t test. *P < 0.05. The exact P values were P = 0.0498 in C; 0.0347 in D; 0.0409 in G; and 0.0437 in H.

Incubation of primary cultures with 10 µM nifedipine to inhibit L-type VGCCs drastically decreased the nuclear Ca²⁺ signals elicited by GBZ (SI Appendix, Fig. S2). Furthermore, the addition of 10 µM GBZ transiently increased the extracellular glutamate levels, detected with the plasma membrane bound fluorescent sensor iGluSnFr expressed in neuronal cells, which exposes its sensor domain to the extracellular medium (SI Appendix, Fig. S3). Albeit the average iGluSnFr fluorescence increase was not statistically significant, the values recorded in each individual culture increased after GBZ addition (SI Appendix, Fig. S3C).

Treatment of primary hippocampal cultures with inhibitory ryanodine suppresses RyR channel activity but preserves ER Ca²⁺ content (44). In addition, inhibitory ryanodine does not affect synaptic transmission in cultured hippocampal slices (45) and preserves paired-pulse facilitation in acute hippocampal slices (46). These previous reports indicate that treatment with inhibitory ryanodine does not affect presynaptic glutamate release. We found that neurons incubated for 1 h with 50 µM ryanodine displayed an initial increase in nuclear Ca²⁺ sensor fluorescence in response to GBZ addition, but the subsequent oscillatory signals had considerably lower magnitude than those displayed by control neurons (Fig. 1B and SI Appendix, Fig. S4 A and B). Following addition of 10 µM gabazine, control neurons displayed higher average values of maximal fluorescence (F_max) in the nucleus (Fig. 1C), higher values of the area under the curve (AUC) (Fig. 1D), and higher oscillatory frequency of the nuclear Ca²⁺ signals compared with the values displayed by neurons treated with inhibitory concentrations of ryanodine (SI Appendix, Fig. S3C).

We have recently reported (34) that transfection of primary hippocampal neurons with shRyR2 (30% efficiency) decreases the total RyR2 protein content of the culture by 40% without modifying RyR3 protein content; in addition, basal RyR1 protein content was very low (SI Appendix, Fig. S5 A and B). Transfection with shRyR2 eliminates agonist-induced RyR-mediated Ca²⁺ release in the transfected neurons, as does 1 h preincubation of control pyramidal neurons with 50 µM ryanodine (SI Appendix, Fig. S5C). Based on these results, we conclude that transfection of primary hippocampal pyramidal neurons with shRyR2 eliminates RyR2 channel activity, which largely mediates Ca²⁺ release induced by the RyR agonist 4-CMC. To test if the RyR2 isoform contributes to the emergence of nuclear Ca²⁺ signals induced by gabazine, primary cultures were cotransfected with GCaMP3.NLS and the same shRyR2 or shScr sequences used in our previous work (34) (SI Appendix, Fig. S1 E and G). Records from a representative experiment show that neurons transfected with shScr displayed synchronic oscillatory Ca²⁺ signals in the nucleus following addition of 10 µM GBZ, with significantly higher nuclear sensor fluorescence as a function of time (Fig. 1E) than that displayed by neurons expressing shRyR2 (Fig. 1F). Neurons transfected with shRyR2 also displayed significantly lower average values of F_max and AUC (Fig. 1 G and H) and lower oscillatory frequency of the nuclear Ca²⁺ signals (SI Appendix, Fig. S4D) in response to the addition of 10 µM GBZ when compared with neurons transfected with shScr.

Stimulation of neuronal activity with uncaged glutamate.

To test the effects of local dendritic stimulation, we uncaged glutamate in dendrites and determined its impact on cytoplasmic and nuclear Ca²⁺ signal generation. In contrast to stimulation with GBZ, which acts globally, glutamate uncaging induces local and time-limited stimulation of a dendrite in a single neuronal cell, excluding possible presynaptic effects or time-delayed changes in the expression of proteins that may be engaged in Ca²⁺ signal propagation to the nucleus. Moreover, axons and presynaptic terminal regions also express RyR channels (47, 48), which may be stimulated by the global stimulation induced by GBZ or HFFS but not by local glutamate uncaging at the dendrites. Glutamate uncaging was done by applying UV laser pulses close to the dendritic surface. This procedure increased glutamate concentration only in the immediate vicinity of the uncaging site, as evidenced by the local increase in the fluorescence of the glutamate sensor iGluSnFr, which is expressed in the neuronal surface (SI Appendix, Fig. S6 A and B and Movie S4). A representative experiment (Movie S5) illustrates that glutamate uncaging in a dendrite, induced by three sequential stimulation protocols of increasing intensity (0.1, 0.5, and 1.0 V, delivered at 31 s, 1.01 min, and 2.02 min), generated cytoplasmic Ca²⁺ signals detected with the GCaMP5 sensor, which propagated bidirectionally along the dendrite and reached the soma around the nucleus.

Next, we tested the effects of glutamate uncaging at the dendrites on nuclear Ca²⁺ signal generation in neurons transfected with GCaMP3.NLS (Fig. 2 A, Left). Images from a representative experiment show that stimulation close to the dendritic neuronal shaft with a series of sequential UV light pulses of increasing intensity to promote glutamate uncaging (arrows, Fig. 2 A, Right) increased Ca²⁺ sensor fluorescence intensity and signal propagation along the dendrite and increased fluorescence intensity at the neuronal nucleus. Increasing the number of pulses produced higher increments; 20 pulses elicited nuclear Ca²⁺ signals that reached intensities comparable to those elicited by spontaneous activity (SA). Quantification of the maximal intensity of the nuclear Ca²⁺ sensor signals, elicited by glutamate uncaging with 10 UV light pulses, shows that shRyR2-transfected neurons displayed significantly lower GCaMP3.NLS fluorescence increases both in the dendrites and the nucleus compared with control neurons (Fig. 2B and SI Appendix, Fig. S7). Sequential stimulation with an increasing number of UV pulses elicited nuclear Ca²⁺ signals that were notably lower in shRyR2 transfected neurons compared with the controls (Movie S6). These results indicate that local glutamate uncaging at the dendrites evokes cytoplasmic Ca²⁺ signals that propagate to the nucleus and that RyR2-mediated Ca²⁺ release plays a significant role in this response.

Fig. 2.

RyR2-mediated Ca²⁺ release is required for the optimal induction of nuclear Ca²⁺ signals in response to glutamate uncaging. (A, Left) A single neuron expressing GCaMP3.NLS was stimulated locally by glutamate uncaging (five pulses of UV light) at the main dendrite, causing an immediate increment in Ca²⁺ sensor fluorescence both at the site of stimulation and at the nucleus. (Scale bar: 30 μm.) Although the GCaMP3.NLS sensor exhibited significantly higher fluorescence in the nucleus, due to its preferential nuclear destination as shown in this image, it also exhibited weak cytoplasmic fluorescence. (Right) Representative pseudoline scan plot of Ca²⁺ signals induced by sequential UV light stimulation (5, 10, and 20 pulses delivered at arrowheads) to elicit glutamate uncaging; images correspond to the same neuron illustrated in the left panel. The generation of dendritic and nuclear Ca²⁺ signals by two events of SA can also be appreciated. (Vertical scale bar: 30 μm; horizontal scale bar: 10 s.) (B) Quantification of the maximal fluorescence intensity [F_max values; expressed as (F_max − F₀)/F₀] elicited at the nucleus by glutamate uncaging at the dendrites; 10 UV light pulses were applied to control (open bar) or to shRyR2-transfected neurons (dark gray bar). All experiments were done using paired control and shRyR2-transfected neurons from the same culture, recorded in the same day 48 h after transfection. Data correspond to mean ± SE; n = 3 to 5 independent cultures (at least 17 neurons were evaluated in each condition). Statistical analysis was performed with the Kolmogorov–Smirnov test; **P < 0.01. The exact P value was 0.008.

Stimulation of neuronal activity with HFFS.

To further investigate the contribution of RyR-mediated Ca²⁺ release to the generation of activity-induced cytoplasmic and nuclear Ca²⁺ signals, HFFS protocols of increasing intensity were applied to primary cultures transfected, respectively, with GCaMP5 or GCaMP3.NLS. Stimulation with HFFS (100 pulses at 20 Hz, delivered at 7 s) generated global cytoplasmic Ca²⁺ signals in both the dendrites and the soma (Movie S7). The sequence of images illustrated in Fig. 3A shows the time-dependent fluorescence changes of the GCaMP3.NLS nuclear sensor displayed by a control neuron stimulated with 200 pulses at 20 Hz. Images from a representative experiment (Fig. 3B) show the sequential increases in nuclear Ca²⁺ sensor fluorescence elicited by sequential stimulation of a single neuron with 100 or 200 pulses at 20 Hz. Stimulation with 400 pulses elicited an even higher increase in nuclear Ca²⁺ sensor fluorescence (SI Appendix, Fig. S8A). Following stimulation with 200 pulses, neurons from cultures preincubated for 1 h with 50 µM ryanodine (Fig. 3C) or transfected with shRyR2 (Fig. 3D) displayed lower increases in nuclear Ca²⁺ signals compared with control neurons or to neurons transfected with shScr, respectively.

Fig. 3.

RyR2-mediated Ca²⁺ release contributes to the generation of nuclear Ca²⁺ signals elicited by HFFS of primary hippocampal neurons. (A) Time-lapse images recorded from a representative experiment showing nuclear fluorescence changes with time displayed by a control hippocampal neuron expressing GCaMP3.NLS and recorded before and after field stimulation with 200 pulses at 20 Hz. (Scale bar: 30 µm.) (B) Representative pseudoline scan plot showing nuclear Ca²⁺ transients recorded in a control neuron expressing GCaMP3.NLS after sequential HFFS with 100 and 200 pulses at 20 Hz. A color-coded nonlinear calibration bar of fluorescence intensity changes is shown at the right of the plot. (Vertical scale bar: 30 μm; horizontal scale bar: 10 s.) (C) Nuclear Ca²⁺ signals were measured in neurons expressing GCaMP3.NLS after sequential HFFS with 50 and 100 pulses at 10 Hz or 100 and 200 pulses at 20 Hz. Fluorescence changes were determined in neurons from control cultures (empty bars) and in neurons present in cultures preincubated for 1 h with 50 µM ryanodine (gray bars); n = 7. (D) Similar determinations as described in C were performed in neurons cotransfected with shScr (empty bars) or shRyR2 (dark gray bars). Data are expressed as mean ± SE; n = 7 (36 to 44 cells per condition). Statistically significant differences among experimental conditions were evaluated by one-way ANOVA followed by Bonferroni’s multiple comparison test; *P < 0.05, **P < 0.01. From left to right, the exact P values in C were 0.0422, 0.0048, 0.0079, and 0.0256. For D, these values were 0.0327 and 0.0271.

Sequential stimulation with two HFFS protocols of increasing intensity elicited significantly lower nuclear Ca²⁺ signals in neurons preincubated with 50 µM ryanodine compared with control neurons (Movie S8); these HFFS protocols also generated lower nuclear Ca²⁺ signals in neurons transfected with shRyR2 relative to neurons transfected with shScr (Movie S9). Compared with their respective controls, the AUC values derived from the nuclear Ca²⁺ signals elicited by different HFFS protocols were significantly lower in neurons from cultures preincubated with 50 µM ryanodine (SI Appendix, Fig. S8B) or transfected with shRyR2 (SI Appendix, Fig. S8C).

The combined results presented in this section show that RyR2-mediated Ca²⁺ release plays a significant role in the generation of nuclear Ca²⁺ signals elicited by local glutamate uncaging or by global stimulation with GBZ or HFFS.

Activity-Generated Gene Expression Changes.

The nuclear increases in free Ca²⁺ concentration elicit the activity of transcriptional regulators, such as CREB and Npas4, and promote the expression of several neuronal genes implicated in synaptic plasticity, learning, and memory, including brain-derived neurotrophic factor (BDNF) (2). Here, we found that addition of 5 µM GBZ to control cultures caused a time-dependent increase in the levels of phosphorylated CREB (P-CREB) in neuronal cells (SI Appendix, Fig. S9). Compared with controls, neurons in cultures preincubated with 50 µM ryanodine displayed a significant decrease in neuronal P-CREB levels after incubation with 5 µM GBZ for 2 h (Fig. 4 A and B).

Fig. 4.

GBZ increases CREB phosphorylation and enhances the expression of Npas4 and RyR2 in primary hippocampal neurons. (A) Cultures were treated with 5 µM GBZ for 2 h, fixed, and immunostained with antibodies for P-CREB (green) and the neuronal marker MAP2 (red); the nucleus was stained with DAPI (blue). Representative images from control hippocampal cultures and from cultures preincubated for 1 h with 50 µM ryanodine before GBZ addition. Images were acquired by spinning disk microscopy, using a 40× objective. (B) Quantification of the relative P-CREB fluorescence levels displayed by three different hippocampal cultures in each condition. Values represent mean ± SE. (C) Relative Npas4 mRNA levels determined by qPCR, normalized to β-actin mRNA levels, and expressed as fold over control. (D) Relative RyR2 mRNA levels determined by qPCR, normalized to β-actin mRNA levels and expressed as fold over control. In C and D, values represent mean ± SE; n = 3 to 5. Statistically significant differences among experimental conditions were evaluated by one-way ANOVA followed by Bonferroni’s multiple comparison test. *P < 0.05; **P < 0.01, ***P < 0.001. The exact P values were 0.01478 for B; 1.43 × 10⁻⁸, 0.00232, and 0.01478 for C; and 0.01308, 0.01293, and 0.03654 for D.

In addition, we found that neurons in primary cultures incubated with 5 µM GBZ for 2 h displayed significant increases in Npas4 (Fig. 4C) and RyR2 (Fig. 4D) messenger RNA (mRNA) levels. To test if RyR-mediated Ca²⁺ release participated in these increments, we preincubated cultures for 1 h with 50 µM ryanodine before GBZ addition. We found that suppression of RyR channel activity caused a significant reduction in the increase in Npas4 and RyR2 mRNA levels induced by GBZ (Fig. 4 C and D). These findings add to previous reports, showing enhanced expression of Npas4 (49, 50) and RyR2 (41, 42, 46, 51) in response to hippocampal neuronal stimulation with different protocols.

Discussion

Hippocampal cultures contain inhibitory interneurons (∼10 to 15%), which impose tonic inhibition on the neuronal network; hence, blocking GABA_A receptor function with bicuculline or GBZ causes neurons to fire synchronous bursts of action potentials (52). Therefore, the effects of bicuculline or GBZ on the generation of nuclear Ca²⁺ signals in pyramidal neurons are indirect. Previous studies have reported that bicuculline addition to primary hippocampal neurons induces the influx of Ca²⁺ through synaptic NMDA receptors and VGCCs (18); the ensuing increase in cytoplasmic Ca²⁺ concentration promotes CICR from intracellular Ca²⁺ stores leading to the subsequent amplification of synaptic Ca²⁺ signals and resulting in Ca²⁺ wave propagation into the nucleus (29). We have confirmed these findings using GBZ to generate nuclear Ca²⁺ transients in primary hippocampal cultures. However, different results have been reported in acute hippocampal slices exposed to trains of synaptic stimuli (high-frequency or theta burst stimulation), in which the nuclear Ca²⁺ signals measured in CA1 pyramidal neurons persist despite SERCA pump inhibition, a condition that prevents Ca²⁺ release by emptying the ER of Ca²⁺ (53). The mechanisms underlying these different responses remain to be explored.

In the present work, we found that GBZ addition promoted glutamate release detected with a fluorescent glutamate sensor expressed in neuronal cells. The present results add to a previous report showing that bicuculline addition to mixed neuron-astrocyte cultures, in which the astrocytes expressed the glutamate sensor, induced glutamate release (54). Also, we present here the finding that RyR2-mediated Ca²⁺ release plays a significant role in the generation of nuclear Ca²⁺ transients induced by GBZ. In addition, we found that local neuronal stimulation by direct uncaging of glutamate in the dendrites or global stimulation by HFFS protocols generated nuclear Ca²⁺ transients that were significantly reduced by RyR inhibition with ryanodine or by RyR2 down-regulation with shRyR2. Accordingly, we propose that RyR2-mediated Ca²⁺ release makes a significant contribution to neuronal nuclear Ca²⁺ signal generation in primary hippocampal cultures.

The rat hippocampus expresses mainly the RyR2 channel isoform and, in lower proportions, the RyR3 and IP₃R1 channel isoforms (36), with very low levels of RyR1 expression (SI Appendix, Fig. S5A). Recent reports have highlighted the pivotal role of RyR2-mediated Ca²⁺ release in dendritic spine remodeling and hippocampal-dependent memory processes (34). Both in vitro and in vivo experiments have confirmed the central role RyR2 channels play in the maintenance and remodeling of dendritic spines, since their loss results in compensatory increases in neuronal excitability and hyperactivity (35). Moreover, neuronal RyR2 channels interact with the L-type Cav1.3 channels; this structural and functional cluster was proposed to have a role in translating synaptic activity into gene expression changes (51). In addition, RyR2 channels form a tripartite complex with the plasma membrane channels CaV1 and KCa3.1, which controls neuronal excitability and the slow posthyperpolarization response (55). These combined results highlight the important role of RyR2 channels in neuronal function.

In addition to the contribution of RyR2-mediated Ca²⁺ release to nuclear Ca²⁺ signal generation reported here, the contributions of IP₃R to this essential neuronal response remain to be explored. It has been reported that hippocampal RyR and IP₃R channels work independently, with separated sources of reticular Ca²⁺ (56). However, suppression of RyR channel activity decreases both the magnitude and propagation of Ca²⁺ signals induced by IP₃ uncaging in dendrites of primary hippocampal neurons, showing functional interactions between these two ER-resident Ca²⁺ channel types (57). Store-operated Orai1 channels amplify glutamate-evoked Ca²⁺ signals in the dendritic spines of hippocampal neurons (58), but their contribution to the propagation of Ca²⁺ signals to the nucleus remains unreported. We did not address here whether IP₃R, RyR3, and Orai1 channels participate in the propagation to the nucleus of activity generated nuclear Ca²⁺ transients. Future studies should address their possible participation in the neuronal excitation–transcription coupling process; this is an important topic, since nuclear Ca²⁺-mediated signaling connects neuronal activity with transcriptional responses that are crucial for the adaptation of neuronal cells to new and varying conditions (2, 3, 8, 23, 59, 60). In addition, sequential activation of NMDARs and L-type VGCCs in hippocampal neurons induces ER Ca²⁺ release, which leads to activation of the stromal interaction molecule 1 (STIM1), an ER resident luminal Ca²⁺ sensor; STIM1 activation, in turn, inhibits L-type VGCCs, and this inhibition regulates both dendritic spine ultrastructure and signaling from the synapse to the nucleus (61). As pointed out by these authors, it remains to be determined whether CICR mediated by RyR and/or IP₃R channels is involved in this response.

Activity-dependent gene expression is closely linked to Ca²⁺ signal generation. The variation in their amplitude and spatiotemporal profiles allows Ca²⁺ signals to trigger different transcriptional responses (37, 62, 63). Pioneer work using microinjection of nondiffusible Ca²⁺chelators into the nucleus revealed that nuclear Ca²⁺ signals regulate gene expression (29, 64). An expanded set of transcription factors and genes, which are functionally associated with acquired neuroprotection, neuronal architecture, and pain, and which are regulated by nuclear Ca²⁺ signals, was established (6, 38, 65, 66). The immediate early genes represent the first line of transcriptional factors that initiate and amplify the transcriptional responses, acting as differential markers for the functional response to different stimulation types that trigger synaptic activity (59). Here, we report that GBZ promoted CREB phosphorylation and enhanced Npas4 and RyR2 expression and that suppressing RyR activity significantly reduced these responses. Our results bring evidence on the contribution of RyR-mediated Ca²⁺ release to the activation of transcriptional factors such as CREB and Npas4. In particular, the induction of Npas4 is one of the most rapid and specific early events that occur after stimulation of neuronal activity (12); RyR channels may contribute to Npas4 activation following neuronal stimulation, since they form part of a molecular complex with VGCCs in microdomains (55, 67). Moreover, Npas4 phosphorylation by MAP kinases increases the interaction of Npas4 with CBP in the mouse striatum, enhancing Npas4 transcriptional activity (14). It remains to be determined if RyR-mediated Ca²⁺ release, which promotes MAP kinase (ERK1/2) activation in primary hippocampal neurons (68), enhances Npas4 phosphorylation and thus promotes Npas4-dependent BDNF expression in hippocampal neurons.

Up-regulation of RyR2 channels occurs following neuronal stimulation. Spatial learning (30, 69) or sustained long-term potentiation induction and hippocampal-dependent recognition memory training (46) significantly increase RyR2 protein levels in rat hippocampus or hippocampal slices. Likewise, HFFS of primary hippocampal cultures promotes RyR2 expression through RyR-mediated CICR activation (41), whereas RyR2 channels mediate the increase in RyR2 protein content induced by BDNF (34). In addition, nicotine administration to mice induces CREB-mediated RyR2 up-regulation in brain areas associated with cognition and addiction; this up-regulation reinforces, via CICR, long-lasting Ca²⁺ signaling and promotes long-term CREB phosphorylation in a positive-feedback signaling loop (70). Based on these previous reports, plus the present results showing that RyR-mediated Ca²⁺ release contributes to the increased CREB phosphorylation and the RyR2 up-regulation induced by neuronal activation, we propose that activation of RyR2-mediated Ca²⁺ release induces RyR2 expression. This proposed positive-feedback mechanism has physiological relevance because inhibition of RyR2-mediated CICR has deleterious effects on hippocampal synaptic plasticity and memory processes (34, 35).

Several issues remain to be addressed in futures studies, including the identification of the various Ca²⁺ signaling pathways by which CICR, mediated by RyR2 and/or IP₃R1 channels, participates in hippocampal synaptic plasticity and spatial memory processes. It also remains to be tested whether RyR2 down-regulation modifies the expression of glutamate receptors or of VGCCs, including L-type VGCCs, among other proteins engaged in Ca²⁺ homeostasis and signaling.

Conclusions

Our results conclusively show that RyR2-mediated Ca²⁺ release contributes to the generation of nuclear Ca²⁺ signals induced by local or global stimulation protocols. In addition, the nuclear Ca²⁺ transients induced by RyR-mediated Ca²⁺ release contribute to enhance CREB phosphorylation and enhance the expression of the neuronal exclusive Npas4 transcription factor and of RyR2 channels. These results further implicate RyR-mediated Ca²⁺ release as a central participant in hippocampal neuronal function.

Data Availability

All data, documentation, and code used in the analysis have been deposited in the Open Science Framework and are included in the following link: https://osf.io/7fd9b/?view_only=d36e8e6b67dd4f0ebc5a186fc99e82a4.