Bidirectional influence of sodium channel activation on NMDA receptor–dependent cerebrocortical neuron structural plasticity

Abstract

Neuronal activity regulates brain development and synaptic plasticity through N-methyl-d-aspartate receptors (NMDARs) and calcium-dependent signaling pathways. Intracellular sodium ([Na⁺]_i) also exerts a regulatory influence on NMDAR channel activity, and [Na⁺]_i may, therefore, function as a signaling molecule. In an attempt to mimic the influence of neuronal activity on synaptic plasticity, we used brevetoxin-2 (PbTx-2), a voltage-gated sodium channel (VGSC) gating modifier, to manipulate [Na⁺]_i in cerebrocortical neurons. The acute application of PbTx-2 produced concentration-dependent increments in both intracellular [Na⁺] and [Ca²⁺]. Pharmacological evaluation showed that PbTx-2–induced Ca²⁺ influx primarily involved VGSC activation and NMDAR-mediated entry. Additionally, PbTx-2 robustly potentiated NMDA-induced Ca²⁺ influx. PbTx-2–exposed neurons showed enhanced neurite outgrowth, increased dendritic arbor complexity, and increased dendritic filopodia density. The appearance of spontaneous calcium oscillations, reflecting synchronous neuronal activity, was accelerated by PbTx-2 treatment. Parallel to this response, PbTx-2 increased cerebrocortical neuron synaptic density. PbTx-2 stimulation of neurite outgrowth, dendritic arborization, and synaptogenesis all exhibited bidirectional concentration–response profiles. This profile paralleled that of NMDA, which also produced bidirectional concentration–response profiles for neurite outgrowth and synaptogenesis. These data are consistent with the hypothesis that PbTx-2–enhanced neuronal plasticity involves NMDAR-dependent signaling. Our results demonstrate that PbTx-2 mimics activity-dependent neuronal structural plasticity in cerebrocortical neurons through an increase in [Na⁺]_i, up-regulation of NMDAR function, and engagement of downstream Ca²⁺-dependent signaling pathways. These data suggest that VGSC gating modifiers may represent a pharmacologic strategy to regulate neuronal plasticity through NMDAR-dependent mechanisms.

Development of the nervous system involves both a predetermined genetic program that controls the general organization of the brain and activity-dependent mechanisms that modulate several developmental processes, including dendritic elaboration, spine formation, synapse formation, and synapse elimination (1). Neuronal activity is one of the essential extracellular signals that influences dendritic arborization (2). During development, dendrites are first decorated by filopodia (3) that are considered to be precursors of spines (4, 5). During initial development, synapses emerge at contacts between axons and dendritic filopodia (6, 7) that may mature into excitatory spine synapses in later development. Previous studies have demonstrated that neuronal activity and activity-dependent gene transcription have profound effects on synaptogenesis (8).

The N-methyl-d-aspartate (NMDA) type of glutamate receptor mediates many of the effects of neuronal activity on structural plasticity. The development and structural remodeling of dendrites and spines depend on actin cytoskeletal reorganization, and cytoskeletal dynamics are regulated by calcium influx through NMDA receptors (NMDARs) and downstream signaling pathways (9, 10). Previous studies have indicated that changes in intracellular sodium concentration ([Na⁺]_i) produced in soma and dendrites as a result of neuronal activity may play a role in activity-dependent synaptic plasticity. Synaptic stimulation causes [Na⁺]_i increments of 10 mM in dendrites and of up to 35–40 mM in dendritic spines (11). In hippocampal neurons, intracellular [Na⁺] increments greater than 5–10 mM have been demonstrated to increase NMDAR-mediated whole-cell currents and single-channel activity by increasing both channel open probability and mean open time (12).

In the present study, we used brevetoxin-2 (PbTx-2), a voltage-gated sodium channel (VGSC) gating modifier, to manipulate [Na⁺]_i in immature cerebrocortical neurons in an effort to mimic neuronal activity. Brevetoxins interact with neurotoxin site 5 on the α subunit of VGSCs (13, 14) and augment sodium influx by shifting the activation potential to more negative values and inhibiting channel inactivation (15). We have shown previously that in cerebrocortical neurons, PbTx-2 elevates [Na⁺]_i and augments NMDAR whole-cell currents by increasing both channel open probability and mean open time without affecting the resting membrane potential (16). PbTx-2 treatment affects neurite outgrowth in day in vitro (DIV)-1 cerebrocortical neurons in a bidirectional pattern (16). We hypothesized that the effects of PbTx-2 on structural plasticity in cerebrocortical neurons may be mediated by NMDAR-dependent mechanisms subsequent to an increase in [Na⁺]_i. Here, we show that acute application of PbTx-2 causes Na⁺ and Ca²⁺ influx, with the [Na⁺]_i increment being sufficient to up-regulate NMDAR function. Furthermore, we demonstrate that PbTx-2 treatment enhances dendritic arborization, filopodia formation, and synaptogenesis. We also demonstrate that PbTx-2 exposure engages downstream activity-dependent mechanisms involving Ca²⁺-calmodulin kinases (CaMKs), ERK, cAMP response element binding protein (CREB), and BDNF signaling pathways. Taken together, our results indicate that PbTx-2 mimics activity-dependent synaptic plasticity by potentiating NMDAR function with downstream engagement of Ca²⁺-dependent signaling mechanisms.

Results

PbTx-2, a Sodium Channel Gating Modifier, Increases Intracellular Sodium and Calcium Influx in DIV-1 Cerebrocortical Neurons.

To determine the extent to which PbTx-2 can elevate [Na⁺]_i in immature DIV-1 cerebrocortical neurons, we imaged individual neurons loaded with the Na⁺-sensitive fluorescent dye sodium-binding benzofuran isophthalate-acetoxymethyl ester (SBFI-AM). As depicted in Fig. 1A, PbTx-2 produced a concentration-dependent elevation of [Na⁺]_i. The PbTx-2 concentration–response data revealed an EC₅₀ of 12.7 nM [95% confidence interval (CI), 5–33 nM] (Fig. 1A). An in situ SBFI calibration (Fig. S1) revealed a basal [Na⁺]_i of 9 ± 3 mM in DIV-1 neurons, and PbTx-2 (1 µM) produced a maximum [Na⁺]_i of 60 ± 10 mM. To confirm the involvement of VGSCs in the observed elevation of [Na⁺]_i, we treated neurons with tetrodotoxin (TTX), a VGSC pore blocker. TTX (1 µM) treatment completely blocked the 100 nM PbTx-2–induced elevation of [Na⁺]_i, establishing VGSCs as the source of Na⁺ entry (Fig. 1B).

Fig. 1.

PbTx-2, a sodium channel gating modifier, increases intracellular sodium and calcium levels in DIV-1 cerebrocortical neurons. (A) PbTx-2 concentration-dependent elevation of [Na⁺]_i. SBFI fluorescence data are expressed as maximum changes in (340/380) ratio divided by the baseline ratio (ΔR/R₀). Right ordinate depicts the calibrated [Na⁺]_i corresponding to SBFI fluorescence ratio. Data shown are from a representative experiment (n = 20–30 cells) repeated three times. (B) Pretreatment with TTX (1 µM) blocked 100 nM PbTx-2–induced increase in SBFI fluorescence. TTX and PbTx-2 addition indicated by arrows. A representative trace is shown. (C) Ca²⁺ imaging with Fura-2 AM showed PbTx-2 concentration-dependent increments in [Ca²⁺]_i. Data are expressed as maximum change in fluorescence ratio (340/380) divided by baseline ratio. Right ordinate indicates [Ca²⁺]_i derived from calibration of the Fura-2 fluorescence. PbTx-2 EC₅₀ value: 360 nM (95% CI, 133-969 nM). Data shown are from a representative experiment (n = 20–30 cells) repeated four times. (D) Pharmacological evaluation of 100 nM PbTx-2–induced Ca²⁺ influx examined using TTX (1 µM), D-APV (100 µM), or nifedipine (1 µM). TTX abrogated the response to PbTx-2, whereas D-APV, but not nifedipine, significantly reduced PbTx-2–induced Ca²⁺ influx. Representative traces are shown; data were collected from cells treated with either 100 nM or 1 µM PbTx-2 (n = 20–30 cells per condition per experiment and repeated thrice). (E) PbTx-2 potentiation of NMDA-induced Ca²⁺ influx. NMDA concentration–response effects on Ca²⁺ influx in the absence and presence of 30 nM PbTx-2. Data shown (means ± SEM) are from a representative experiment (n = 30 cells) performed in sister coverslips repeated in three independent cultures.

Given that activity-dependent neuronal development and structural plasticity primarily engage Ca²⁺-dependent signaling pathways (17), we next assessed whether PbTx-2 affects Ca²⁺ dynamics in these neurons. PbTx-2 produced a rapid and concentration-dependent increase in [Ca²⁺]_i (Fig. 1C and Fig. S2). Calibration of Fura-2 fluorescence showed a basal [Ca²⁺]_i of 60 ± 20 nM, and 1 µM PbTx-2 produced a [Ca²⁺]_i of 1,320 ± 413 nM (Fig. 1C). To delineate the Ca²⁺-influx pathways triggered by PbTx-2, a pharmacological evaluation of the [Ca²⁺]_i response to 100 nM PbTx-2 was performed. Neurons were pretreated with the antagonists TTX (VGSC), amino-5-phosphonopentanoic acid (D-APV) (NMDAR), or nifedipine (L-type voltage-gated calcium channel) before PbTx-2 challenge (Fig. 1D). TTX (1 µM) completely blocked the response to PbTx-2, whereas D-APV (100 µM), but not nifedipine (1 µM), significantly reduced PbTx-2–induced Ca²⁺influx. A pharmacological evaluation of PbTx-2 induced Ca²⁺ influx at the population level using neurons loaded with Fluo3-AM yielded similar results (Fig. S3). These experiments indicate that PbTx-2–induced Ca²⁺ influx requires VGSC activation and primary cell entry through the calcium-permeable NMDAR. Given the ability of PbTx-2 exposure to augment NMDAR channel function (12, 16), we next evaluated the ability of PbTx-2 treatment to affect NMDA-induced Ca²⁺ influx at the single-cell level in DIV-1 neurons. NMDA alone produced a concentration-dependent elevation of [Ca²⁺]_i (Fig. 1E) with an EC₅₀ of 444.1 nM (95% CI, 24 nM to 8.2 µM). In the presence of a concentration of PbTx-2 (30 nM) that alone produced a modest transient increase in [Ca²⁺]_i, the NMDA-induced Ca²⁺ influx was robustly potentiated. The NMDA concentration–response curve displayed a leftward shift [NMDA plus PbTx-2: EC₅₀, 12 nM (95% CI, 4.7–30.2 nM)] and an increased maximum response (E_max) in the presence of PbTx-2 [NMDA: E_max = 0.013 (95% CI, 0.011–0.016); NMDA plus PbTx-2: E_max = 0.025 (95% CI, 0.022–0.028)]. These findings confirm the ability of PbTx-2 treatment to enhance NMDAR function in cerebrocortical neurons.

PbTx-2–Enhanced Neurite Outgrowth Exhibits a Bidirectional Profile Similar to that of NMDA.

We have shown previously that the effect of PbTx-2 on neurite outgrowth exhibited a bidirectional concentration–response profile and that this effect was primarily dependent on NMDARs (16). Lipton and Nakanishi (18) have similarly described an inverted-U profile for the relationship between NMDAR activity and neuronal survival, with too little or too much NMDAR activity resulting in diminished growth or neuronal death. We reasoned that PbTx-2 and NMDA may share common downstream signaling pathways and, therefore, directly compared their effects on neurite outgrowth in DIV-1 neurons. The PbTx-2 and NMDA concentration–response curves for neurite outgrowth both exhibited bidirectional profiles consistent with a common NMDAR-dependent mechanism (Fig. 2 A and B).

Fig. 2.

PbT-2–enhanced neurite outgrowth and dendritic arborization exhibit a bidirectional concentration–response profile. (A) Representative images of NMDA-induced neurite outgrowth. (Scale bar: 10 µm.) (B) Quantification of neurite outgrowth 24 h after plating. PbTx-2 or NMDA were added to culture at 3 h postplating. Each value represents mean ± SEM of 120 neurons. (C) Representative images of 3′-tetramethylindocarbocyanine (DiI)-loaded cerebrocortical neurons at DIV-4. (D) Sholl analysis to quantify arbor complexity. (E) Area under the curve (AUC) analysis of Sholl data. PbTx-2 at 30 and 100 nM significantly enhanced dendritic complexity. The experiment was performed thrice, and each point represents the mean derived from the analysis of 30–40 neurons (ANOVA, P < 0.01; Dunnett’s post hoc test, P < 0.05). (F) Pharmacological evaluation of 30 nM PbTx-2–induced dendritic arborization. The response to 30 nM PbTx-2 exposure was examined in the presence of TTX (1 μM), MK-801 (1 μM), or nimodipine (1 μM) beginning at 3 h after plating. Each data point represents mean ± SEM of 50–60 neurons. (G) AUC analysis of the data shown in F. PbTx-2–enhanced dendritic arbor complexity was significantly attenuated by MK-801 and TTX. (ANOVA, P < 0.01; Bonferroni post hoc test, *P < 0.05).

PbTx-2 Enhances Dendritic Arbor Development and Stimulates Filopodia Formation.

NMDARs play a critical role in activity-regulated expansion of dendritic arbors through Ca²⁺-dependent signaling cascades (17, 19). Inasmuch as PbTx-2 augments NMDAR function and enhances neurite outgrowth, we determined the effect of PbTx-2 on the development of dendritic arbors and filopodia. In DIV-4 control neurons, there is a gradual increase in branch complexity moving away from the soma reaching a maximum of 7.5 ± 0.6 intersections per neuron and then progressively declining beginning ∼10 µm from the soma (Fig. 2 C and D). PbTx-2 (30 and 100 nM) produced robust increases in dendritic complexity and a rightward shift in the Sholl plot compared with control neurons (Fig. 2D) (20). An area under the curve (AUC) analysis of these Sholl data showed a significant increase in dendritic complexity following either 30 or 100 nM PbTx-2 compared with control (ANOVA, P < 0.01; Dunnett’s post hoc test, P < 0.05) (Fig. 2E). The PbTx-2 effect on dendritic complexity again displayed a bidirectional profile, with the 300 and 1,000 nM concentrations showing progressively smaller responses (Fig. 2E). Using selective pharmacological inhibitors, we next explored the signaling mechanisms underlying the enhanced dendritic arborization produced by PbTx-2. As depicted in Fig. 2 F and G, coincubation of TTX (1 µM) or (+)-5-methyl-10,11-dihydro-5H-dibenzo[a, d]cyclohepten-5,10-imine maleate (MK-801) (1 µM) with 30 nM PbTx-2 markedly attenuated (ANOVA, P < 0.01; Bonferroni post hoc test, P < 0.05) the enhancement of dendritic complexity. In contrast, nimodipine (1 µM), an L-type calcium channel blocker, did not produce a statistically significant reduction in the enhancement of dendritic arborization produced by PbTx-2 (P > 0.05). These results suggest that PbTx-2 activation of sodium channels with attendant enhancement of NMDAR signaling is responsible for the expansion of dendritic arbors.

On DIV-5, 30 nM PbTx-2–treated neurons displayed a dramatic increase in the density of dendritic protrusions [32.4 ± 3.3 protrusions per 50 µm of dendrite vs. control (9.6 ± 1.1)] (Fig. 3). Morphological analyses of these protrusions with a defined algorithm for length and head diameter showed that most of the protrusions in PbTx-2–treated neurons were long filopodial-like structures (Fig. 3 A and B). On DIV-6, PbTx-2–exposed neurons revealed a shift in the distribution of protrusions from predominantly long filopodial-type (12.5 ± 3.04) to stubby filopodial-type protrusions (20 ± 3.5) (Fig. 3 C and D). Elongated filopodia are referred to as immature or exploratory-type filopodia principally looking for contact partners, and stubby/club-shaped filopodia are referred to as mature filopodia (3). This indicates that PbTx-2 enhanced both the formation and maturation of filopodia in cerebrocortical neurons. Blocking neuronal activity by acute treatment with TTX has been shown to shift mature filopodia to exploratory filopodia, indicating that filopodia maturation is a process dependent on neuronal activity (3). TTX and APV treatment, but not nimodipine, antagonized the PbTx-2–induced dendritic arbor complexity and increase in filopodial density (Fig. 3A). These data suggest that the PbTx-2 influence on NMDAR signaling accelerates the development of cerebrocortical neurons by stimulating dendritic arbor elaboration and both filopodia formation and maturation.

Fig. 3.

PbTx-2 enhances filopodia formation. (A) Representative images of DiI-labeled neurons at DIV-5. Depicted are PbTx-2 concentration-dependent effects on protrusion density (protrusions/50 µm neurite length) and pharmacological evaluation of PbTx-2–induced (30 nM) protrusion density. (Scale bar: 5 µm.) (B and D) Representative images and analysis of long and stubby filopodial protrusions. PbTx-2 significantly enhanced total protrusion density on DIV-5 and -6 compared with control (^δP < 0.05, Student’s t test), and on DIV-6, PbTx-2–exposed neurons showed a significant shift from predominantly long to stubby filopodial protrusions (*P < 0.05, Student’s t test). Each bar represents the mean ± SEM of 20–30 neurons. (C) Representative images of Dil-labeled neurons at DIV-6 showing PbTx-2 (30 and 100 nM) enhancement of filopodial maturation.

PbTx-2 Enhances Neuronal Connectivity As Revealed by an Acceleration of the Appearance of Spontaneous Calcium Oscillations and Synapse Formation.

Next, we examined the influence of PbTx-2 on spontaneous Ca²⁺ oscillations that report the degree of neuronal connectivity in a population of cerebrocortical neurons. Spontaneous synchronized depolarization, detectable as intracellular calcium oscillations, reflect typical network activity characteristic of cerebrocortical neuron cultures (21). These spontaneous Ca²⁺oscillations become prominent from DIV-6 onward in both amplitude [3,391 ± 173.4 fluorescence units (FU)] and frequency (11 ± 1.5/300 s) (Fig. S4). The presence of spontaneous Ca²⁺ oscillations was assessed from DIV-2 onward at 12-h intervals (Fig. 4). PbTx-2 (30 nM) accelerated the emergence of spontaneous Ca²⁺ oscillations by 24 h (DIV-5), shifting leftward the Ca²⁺ oscillation time–response curves for both amplitude and frequency (Fig. 4 B and C). As depicted in Fig. 4A, control Ca²⁺ oscillations did not become prominent until 148 h (DIV-6), at which time, the oscillation amplitudes reached values of 3,000 FU. In contrast, PbTx-2–exposed neurons displayed greater oscillation amplitudes 24 h earlier at 124 h (DIV-5) (124 h: 4,209.6 ± 692 FU). Analysis of oscillation frequency (number of spikes/300 s) in the PbTx-2–treated cultures also showed a significant increase from 124 h onward compared with control [control vs. PbTx-2: 4.8 ± 2 vs. 10.5 ± 1.8 (124 h), 4.8 ± 2.2 vs. 13 ± 2.5 (136 h), and 9 ± 1.1 vs. 15.8 ± 2.7 (148 h)] (Fig. 4C).

Fig. 4.

PbTx-2 exposure accelerates the appearance of spontaneous Ca²⁺ oscillations. (A) Effect of PbTx-2 (30 nM) on Ca²⁺ oscillations. PbTx-2 was added in alternate wells of 96-well plates beginning 3 h postplating, and plates were scanned for Ca²⁺ oscillations at the indicated time points. Data are from a representative experiment performed in quadruplicate and repeated twice. (B and C) Quantification of data shown in A. (B and C) Mean ± SEM of peak amplitude (B) and frequency (C) of Ca²⁺ oscillations (number of spikes per 300 s). PbTx-2 accelerated the appearance of spontaneous Ca²⁺ oscillations by 24 h (DIV-5) and produced a significant leftward shift in the time–response relationship for both oscillation amplitude and frequency (*P < 0.05, Student’s t test comparison vs. control at respective time points).

Given the strong association between spontaneous calcium oscillations and maturation of synapses and pattern-generating circuits, we reasoned that the accelerated appearance of spontaneous calcium oscillations produced by PbTx-2 may reflect an increased rate of synaptogenesis. Considering that synchronized Ca²⁺ oscillations represent a functional assay for synaptogenesis, we directly assessed the effect of PbTx-2 on synapse density in cerebrocortical neuron cultures. Antibodies against synaptophysin (presynaptic marker) and postsynaptic density 95 (PSD-95) (postsynaptic marker) were used to quantify synapse density, as indicated by colocalized fluorescent puncta. We exposed neurons to PbTx-2 beginning 3 h postplating, and synapse density was assayed on DIV-5. Concentration–response analysis of the influence of PbTx-2 on synaptic density indicated that PbTx-2 at concentrations of 30 and 100 nM produced a significant (ANOVA, P < 0.01; Dunnett’s post hoc test, P < 0.05) increase in synapse formation (number of puncta per 20 µm) compared with control cultures. Separate analysis of pre- and postsynaptic puncta showed an increase in PSD-95, but not synaptophysin, puncta following PbTx-2 exposure. Higher concentrations of PbTx-2 (300 and 1,000 nM) produced a progressive decline in synaptic density [control, 0.95 ± 0.2 puncta per 20 μm of dendrite; PbTx-2 (30 nM), 3.99 ± 0.24; PbTx-2 (100 nM), 3.8 ± 0.34; PbTx-2 (300 nM), 2.53 ± 0.24; PbTx-2 (1,000 nM), 0.79 ± 0.27] (Fig. 5 A and B). We theorized that if the influence of PbTx-2 on synaptic density was the result of augmentation of NMDAR function, then NMDA should also display a bidirectional concentration–response profile on synaptic density. Consistent with this hypothesis, NMDA produced a bidirectional concentration–response profile for synaptogenesis (Fig. 5 C and D), suggesting that common NMDAR-dependent mechanisms are operative in these effects of PbTx-2 and NMDA.

Fig. 5.

Effect of PbTx-2 on formation of synapses in cerebrocortical cultures. (A) Representative images of double-immunostained cerebrocortical neurons at DIV-5. PbTx-2 was added to cultures 3 h after plating. Antibodies against synaptophysin (presynaptic marker/green) and PSD-95 (postsynaptic marker/red) were used to quantify synapse density, as indicated by colocalized fluorescent puncta (yellow). Boxed regions are shown in higher magnification separately and overlaid. (Scale bar: 2 µm.) (B) Quantification of colocalized fluorescent puncta (each bar represents n = 3 independent experiments with duplicate samples; ANOVA, P < 0.01; Dunnett’s post hoc test, *P < 0.05). (C and D) Representative images and quantification of NMDA concentration–response effect on synaptogenesis (n = 3 independent experiments with duplicate samples; ANOVA, P < 0.01; Dunnett’s post hoc test, *P < 0.05). (Scale bar: 2 µm.)

PbTx-2 Exposure Engages Activity-Dependent Signaling Mechanisms That Drive Structural Plasticity.

Inasmuch as activity-dependent neuronal development primarily involves Ca²⁺-mediated signaling events including CaMKs downstream from NMDARs, we have examined whether the increase in [Ca²⁺]_i evoked by PbTx-2 also engages these signaling pathways. Acute exposure to PbTx-2 (100 nM) does produce a rapid and sustained phosphorylation of CaMK1 (Thr177) and CaMKII (Thr286), whereas stimulation of phospho-ERK 1/2 is phasic (Fig. S5 A and B). In contrast, PbTx-2 treatment does not activate either CaMKIV (Thr196) or Akt (Ser473). The transcription factor CREB has been implicated in activity-dependent neuronal development, where it is regulated by CaMKs and MAPKs (22). PbTx-2 exposure produces a marked increase in phosphorylation of CREB at Ser133, and this response is abrogated by TTX (1 µM), D-APV (100 µM), or the CaMKK inhibitor 1,8-naphthoylene benzimidazole-3-carboxylic acid (STO609) (2.6 µM) (P < 0.05). However, nimodipine (1 µM); 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-2,3-dione (NBQX) (5 µM), an AMPA receptor blocker; and ω-conotoxin (1 µM), an N-type Ca²⁺ channel blocker, all were without effect on PbTx-2–induced activation of CREB (Fig. S5 C and D). These data are consistent with a previous study (23) showing that Ca²⁺ influx through NMDARs, rather than L-type VGCCs, leads to robust CREB (Ser133) phosphorylation in neurons. Our results indicate that NMDAR-mediated Ca²⁺ influx is required for PbTx-2–induced CREB phosphorylation. BDNF is an activity-dependent trophic factor that mediates many aspects of neuronal plasticity and is regulated by CREB-dependent transcription (24). The activity-dependent transcription of BDNF mRNA is a rapid response (25). We, therefore, assessed the influence of PbTx-2 on BDNF transcription with conventional RT-PCR in DIV-1 cerebrocortical neurons. BDNF mRNA expression level increased in a time- and concentration-dependent manner following PbTx-2 exposure (Fig. S5E).

The BDNF-tropomyosin-related kinase B (TrkB) receptor signaling complex exerts neurotrophic effects in both the developing and mature nervous system, and TrkB receptor surface expression is regulated by neuronal activity (26). Using a biotinylation assay in DIV-1 cerebrocortical neurons, we have found that PbTx-2 exposure affects surface expression of TrkB receptors. Cell surface expression of TrkB receptors (control, 23.6 ± 3.4% of total) was increased substantially within 30 min of PbTx-2 exposure (PbTx-2: 30 nM, 62.9 ± 7.8%; 100 nM, 40.3 ± 10.6%; 1 µM, 62.7 ± 14.7% compared with the control fraction; Fig. S5 F and G). Given the predominance of GluN2B subunit–containing NMDARs in immature neurons (27) and the role of GluN2B subunits in the regulation of NMDAR cell surface expression (28), we also evaluated the influence of PbTx-2 on GluN2B cell surface expression. The control level of GluN2B subunit surface expression was 18.1 ± 5.25%, and PbTx-2 exposure significantly increased the surface expression of GluN2B subunits (PbTx-2: 30 nM, 26.7 ± 4.35%; 100 nM, 39.9 ± 1.2%; 1 µM, 49.2 ± 10.02% compared with control; Fig S5 F and G). PbTx-2 exposure, therefore, acutely increases both TrkB and NMDAR surface expression in cerebrocortical neurons.

Discussion

Na⁺ influx is fundamental to electrical signaling in the nervous system and is essential for the upstroke of action potentials (29). In murine neurons, VGSCs appear as early as embryonic day 12 (30). Given the role of VGSCs in controlling action potential generation, they serve as candidates for mediating forms of synaptic plasticity that depend on neuronal firing (31). Neuronal activity exerts profound effects on brain development and the structural plasticity of dendritic arbors and spines (17, 32). Toward the goal of mimicking activity-dependent phenomena, we used a VGSC gating modifier to elevate neuronal [Na⁺]_i and potentiate NMDAR function (12, 16). An interesting feature of our results is that the influence of PbTx-2 on neuronal structural plasticity followed a bidirectional pattern similar to that of NMDA. This result is consistent with a role for [Na⁺]_i in enhancing NMDAR function with engagement of downstream mechanisms involved in the regulation of neuronal plasticity.

Previous reports suggested that [Na⁺]_i may act as a signaling molecule to influence NMDAR function through Src-induced phosphorylation of the receptor (12). High firing rates that are capable of inducing long-term potentiation (LTP) are associated with elevated [Na⁺]_i and a Src-induced up-regulation of NMDAR function (13). To unambiguously confirm the enhancement of NMDAR function by PbTx-2, we have previously recorded single-channel currents from cell-attached patches on cerebrocortical neurons; PbTx-2 treatment increased both the mean open time and open probability of NMDARs in the absence of an influence on resting membrane potential (16). These results were extended in the present study by showing that PbTx-2 treatment enhanced NMDA-induced Ca²⁺ influx in cerebrocortical neurons. These findings are consistent with a regulatory influence of Na⁺ on NMDAR function.

NMDAR stimulation promotes dendritic arbor growth, whereas pharmacological blockade of NMDARs reduces dendritic growth (33). The NMDA-induced increase in dendritic arbor complexity is suppressed by blockers of voltage-gated calcium channels or inhibitors of intracellular calcium mobilization (34), indicating that NMDAR activation recruits Ca²⁺ signaling pathways to regulate dendritic growth (19). Consistent with these reports, we found that exposure to PbTx-2 increased neuronal [Ca²⁺]_i, with NMDAR channels constituting the primary source of Ca²⁺ influx in cerebrocortical neurons.

The effects of exposure to PbTx-2 on dendritic arbor complexity are consistent with a previous demonstration that neuronal activity affects structural plasticity in vivo through NMDAR-triggered intracellular signaling events (35). In accordance with earlier reports for activity-dependent plasticity, we found that PbTx-2 treatment increased neurite outgrowth, dendritic arborization, and filopodia formation and maturation. The PbTx-2–enhanced dendritic arborization was mediated by NMDAR-dependent mechanisms: consonant with a report of activity-dependent dendritic arborization requiring NMDAR activation in DIV-9 hippocampal neurons (36). The exposure to PbTx-2 also enhanced synaptogenesis. As a consequence of this latter action, the appearance of spontaneous [Ca²⁺]_i oscillations was accelerated. These observations implicate synaptic mechanisms in the observed PbTx-2 acceleration of the development of neuronal connectivity in cerebrocortical neuron cultures.

The effects of PbTx-2 on dendritic arborization and synaptogenesis displayed bidirectional concentration–response profiles. Moreover, PbTx-2 effects on calcium influx, dendritic arborization, and filopodia formation were dependent on NMDARs, underscoring the role of NMDAR signaling in the influence of PbTx-2 on neuronal structural plasticity. Additional evidence for NMDAR involvement in the effects of PbTx-2 on structural plasticity was the similar bidirectional patterns shown by NMDA and PbTx-2 on both neurite outgrowth and synaptogenesis. An inverted-U model describes the relationship between NMDAR activity and neuronal survival and growth (18). This inverted-U NMDA concentration–response relationship has primarily, but not exclusively, been regressed to intracellular Ca²⁺ regulation. An optimal window for [Ca²⁺]_i is required for activity-dependent neurite extension and branching, with low levels stabilizing growth cones and high levels stalling them, in both cases preventing extension (37). Given that the major source of Ca²⁺ influx following PbTx-2 exposure is through NMDAR channels, we suggest that similar mechanisms are operative in the bidirectional effects of NMDA and PbTx-2 on structural plasticity.

Trafficking and surface expression of NMDARs are also regulated by neuronal activity (38) and phosphorylation of the GluN2B subunit (39). The GluN2B subunit is a primary determinant of the surface mobility of NMDARs (28), and acute PbTx-2 exposure was found to enhance the surface expression of this subunit. PbTx-2 was also capable of activating neuronal activity-dependent signaling targets such as CaMKs, ERK, and CREB. Activity-dependent increases in intracellular Ca²⁺ have been shown to enhance ERK signaling, with subsequent CREB phosphorylation downstream of NMDARs in cerebrocortical neurons (40). Consistent with this earlier report, we found that PbTx-2 stimulation of CREB phosphorylation was NMDAR-dependent. CREB has been shown to mediate activity-dependent dendrite lengthening (36), and CREB stimulation leads to an activation of BDNF gene expression. BDNF increases dendritic complexity in neurons and blockade of TrkB receptors abrogates this effect (41). The ability of PbTx-2 treatment to increase bdnf mRNA expression and TrkB surface localization demonstrates the influence of neuronal activity on these phenomena. VGSC gating modifiers may accordingly represent a pharmacologic strategy to regulate neuronal plasticity through an NMDAR-dependent mechanism. Recent studies have shown that neuronal rewiring following a stroke event occurs as a result of heightened plasticity in the periinfarct cortex, with enhanced axonal sprouting, synaptogenesis, and turnover of dendritic spines (42). Because the processes involved in rewiring and recovery after stroke are similar to those regulating development, it is feasible that VGSC gating modifiers that augment developmental dendritogenesis and synaptogenesis may also improve stroke recovery.

Materials and Methods

Brevetoxin-2 (PbTx-2) was isolated and purified from Karenia brevis cultures at the Center for Marine Sciences at the University of North Carolina (Wilmington, NC). Fluo-3 AM, SBFI-AM, and Pluronic acid were purchased from Invitrogen.

Cell Culture.

Primary cultures of cerebrocortical neurons were prepared from embryos of Swiss–Webster mice on embryonic day 16 as described (16). The Creighton University Institutional Animal Care and Use Committee approved all animal use protocols.

Calcium and Sodium Imaging.

Intracellular calcium and sodium concentration in neurons were assessed using the respective fluorescent dyes Fura-2 and SBFI. Additional details are provided in SI Text.

Double Immunofluorescence and Image Acquisition and Colocalization Analysis.

Double immunostaining with synaptophysin (green), a presynaptic marker, and PSD-95 (red), a postsynaptic marker, were used to quantify colocalized puncta (yellow) as an indication of the presence of a synapse. Additional details are provided in SI Text.

Diolistic Labeling and Quantification of Dendritic Branching and Filopodial Protrusion Density.

The details of diolistic neuron labeling and confocal imaging acquisition are described in SI Text.