Activation of cyclic GMP-dependent protein kinase blocks alcohol-mediated cell death and calcium disruption in cerebellar granule neurons

Abstract

Alcohol during brain development leads to the widespread neuronal death observed in fetal alcohol spectrum disorders (FASD). In comparison, the mature brain is less vulnerable to alcohol. Studies into maturation-acquired alcohol resistance uncovered a protective mechanism that reduces alcohol-induced neuronal death through nitric oxide-cGMP-cyclic GMP-dependent protein kinase (NO-cGMP-cGK) signaling. However, the downstream processes underlying this neuroprotection remain unclear. Alcohol can disrupt levels of intracellular calcium ([Ca²⁺]_i) in vulnerable neuronal populations to trigger cell death in both in vivo and in vitro models of FASD. Since cGK has been demonstrated to regulate and inhibit intracellular Ca²⁺ release, we examined the hypothesis that cGK confers alcohol resistance by preventing [Ca²⁺]_i disruptions. Alcohol resistance, determined by neuronal survival after 24 hours of alcohol exposure, was examined in primary cerebellar granule neuron (CGN) cultures derived from 5 to 7 day-old neonatal mice with an activator, 8-Br-cGMP, and/or an inhibitor, Rp-8-pCPT-cGMPS, of cGK signaling. Intracellular Ca²⁺ responses to alcohol were measured by ratiometric Ca²⁺ imaging in Fura-2-loaded CGN cultures after 8-Br-cGMP treatment. Our results indicate that activating cGK with 8-Br-cGMP before alcohol administration provided neuroprotection, which the cGK inhibitor, Rp-8-pCPT-cGMPS, blocked. Alcohol exposure elevated [Ca²⁺]_i, whereas 8-Br-cGMP pretreatment reduced both the level of the alcohol-induced rise in [Ca²⁺]_i as well as the number of cells that responded to alcohol by increasing [Ca²⁺]_i. These findings associate alcohol resistance, mediated by cGK signaling, to reduction of the persistent and toxic increase in [Ca²⁺]_i from alcohol exposure.

Introduction

Fetal Alcohol Spectrum Disorders (FASD) describes a continuum of permanent birth defects triggered by prenatal alcohol (ethanol) exposure and remains the single leading preventable cause of mental retardation [1]. In comparison, the adult brain is less prone to alcohol toxicity, suggesting that neuroprotective mechanisms exist in the mature brain to protect it from alcohol.

Maturation-acquired neuronal resistance has been demonstrated in rodent models of FASD. For example, alcohol exposure in mice and rats during postnatal days 4-5, but not later during postnatal days 8-9, severely depletes cerebellar granule and Purkinje neurons, reducing cerebellar size [2-5]. Since development of alcohol resistance coincides with expression of neuronal maturation markers, this suggests that susceptible neuronal populations in the cerebellum acquire an alcohol resistance as they mature [5]. Cerebellar granule neuron (CGN) cultures derived from neonatal mice also show an innate alcohol vulnerability, which diminishes as the neurons mature in culture [6]. Given the devastating CNS injury that fetal alcohol exposure produces, understanding the mechanisms behind this acquired neuronal alcohol resistance is crucial.

We previously established [7-9] that the alcohol resistance of CGN cultures can be enhanced by activating neuronal nitric oxide synthase (nNOS) to increase intracellular nitric oxide (NO) levels, which in turn stimulates soluble guanylyl cyclase (sGC) and cyclic guanosine monophosphate (cGMP)-dependent protein kinase (cGK) downstream. Activation of either sGC or cGK diminishes alcohol-mediated cell death in CGN cultures, while inhibition of either blocks neuroprotection from NO [7-9]. Deficient nNOS expression in neonatal mice [10, 11] exacerbates the alcohol-induced neuronal loss. Since the acquired alcohol resistance in the rat cerebellum also coincides with a developmental increase of nNOS expression [5, 12, 13], these studies indicate that the NO pathway has a physiological role in providing resistance to alcohol.

Recently, our attention has focused on downstream cGK targets that can transmit the neuroprotective signal provided by the NO pathway. Although one downstream effector of neuroprotection has been revealed to be nuclear factor-κB (NF-κB) [14, 15], cGK can influence multiple downstream targets to offer alcohol resistance. Alcohol has been observed in both in vitro [16-19] and in vivo models of FASD [20, 21] to trigger neuronal death by disrupting intracellular Ca²⁺ ([Ca²⁺]_i). Conversely, cGK has been demonstrated to regulate and inhibit [Ca²⁺]_i release in multiple contexts [22, 23], including cerebellar tissue [24].

This evidence has led us to examine the relationship between cGK modulation of [Ca²⁺]_i and alcohol-induced neuronal death. In this study, we tested the hypothesis that cGMP activation of cGK can prevent the alcohol-induced [Ca²⁺]_i elevations and resulting neuronal death. Experiments using alcohol-vulnerable CGN cultures examined 1) neuronal survival and 2) [Ca²⁺]_i following alcohol exposure in the presence of activators and/or inhibitors of cGK signaling. Our findings suggest that cGK mediates alcohol resistance by moderating alcohol’s disruption of [Ca²⁺]_i.

Materials and methods

Animals

A breeding colony of C57BL/6 mice (initially obtained from Jackson Laboratories, Bar Harbor, ME) was housed in The University of Iowa Animal Care Facility. All animal experiments were approved by The University of Iowa Institutional Animal Care and Use Committee and performed according to the guidelines of the National Institutes of Health.

Cell culture

Primary CGN cultures offer a well-defined and manageable biological system widely used to investigate the direct neuronal effects of alcohol [25]. These cultures do not proliferate and any cell number reduction is due to decreased neuronal survival [6].

CGN cultures were established from neonatal mice (5–7 days old) as previously described [18]. For all treatment conditions in a single experiment, cultures derived from each mouse litter were established in duplicate or triplicate and constituted one replicate. CGN were plated into poly-D-lysine (PDL, 50 mg/ml)-coated 96-well culture trays (1.5 × 10⁶ cells/ml; 300 μl/well) for cell survival experiments, and into six-well culture trays (1.0 × 10⁶ cells/ml; 2.0 ml/well) which contained PDL-coated (100 mg/ml) glass coverslips (25 mm diameter) for Ca²⁺ imaging experiments. CGN cultures were incubated overnight at 37°C in humidified air (5% CO₂) prior to use.

Pharmacological treatments and alcohol exposure

Alcohol exposure at 400 mg/dl (87 mM ethanol) was chosen for most experiments which began 24 hours after plating. Although chronic alcoholics can achieve similar blood alcohol concentrations [26], it is a high physiological dose. We previously demonstrated that alcohol elicits Ca²⁺-dependent cell death in a concentration-dependent manner in CGN cultures and that lower alcohol concentrations (100 and 200 mg/dl) produced cell death to a smaller degree [6, 18]. Since this indicates that higher alcohol concentrations are eliciting the same effect with greater magnitude, 400 mg/dl was used to more readily achieve statistical significance. This alcohol concentration also matches blood alcohol levels often attained in rodent studies of alcohol neuroteratology [10, 12, 27-29] and our previous in vitro studies [8, 9, 18, 30].

Cultures were also treated with one or more of the following: PBS (vehicle control); 8-Br-cGMP, a cGMP analogue that activates cGK, or Rp-8-pCPT-cGMPS, a specific inhibitor of cGK. Exposure occurred 30 minutes before adding alcohol as described. Culture trays in cell survival experiments were placed in sealed containers containing alcohol baths of equal concentration (0 or 400 mg/dl) to that of culture media, 5% CO₂, and incubated at 37°C.

Cell survival experiments

Twenty-four hours after adding alcohol, viable cells were counted using a dye-exclusion method (trypan blue is excluded by viable cells). CGN cultures suspended in 0.2% trypan blue, were viewed under an inverted phase-contrast microscope (Diaphot; Nikon Instruments Inc., Melville, NY) and counted on a hemocytometer as previously described [18].

Ca²⁺ measurements

Intracellular [Ca²⁺] was determined using the ratiometric fluorescent indicator, Fura-2 [31] and a digital imaging microscope system (Deltascan-1; Photon Technology International, Birmingham, NJ) as previously described [18].

Statistical analyses

For cell survival experiments, data were expressed as “cell number/well” and as percent alcohol-induced cell loss [18, 30]. For each replicate, the difference in cell number between alcohol-treated and non-alcohol control cultures is expressed as a percent of the non-alcohol control. This accounts for variability in the natural responses of CGN cultures derived from different litters (replicates) and plating conditions. Statistical differences in cell numbers and percent cell loss were analyzed with repeated measures ANOVAs, with alcohol and test agent concentrations considered as variables with repeated measures within each replicate. Post hoc tests consisted of Bonferroni-corrected paired t-tests and pairwise comparisons in cell number and percent cell loss analyses, respectively.

Given variability in Ca²⁺ measurement experiments, some cells (2.2%) display an exceedingly low or high baseline [Ca²⁺]_i (below 10 nM or above 200 nM) prior to alcohol addition and were removed from analyses. A mean [Ca²⁺]_i was calculated at each 5 second time point for each cell in the view-field (39.9 ± 2.1 cells, range of 18 to 50 cells) per coverslip. Cells from experimental replicates were pooled to increase the accuracy of [Ca²⁺]_i calculations. Planned comparisons (Student’s t-test) between treatments were made using the maximum [Ca²⁺]_i level (peak [Ca²⁺]_i) each cell attained during the 9 minute testing period [18]. Percentages of cells with large [Ca²⁺]_i responses (attaining peaks of 200 nM or greater) per treatment were analyzed by chi-square (χ²).

All data are presented as mean ± SEM, with statistical differences of p ≤ 0.05 considered significant, except when Bonferroni-corrected thresholds were employed for multiple comparisons. All statistical analyses were performed using SPSS Statistics (IBM, New York, NY).

Results

Activation of cGK ameliorated alcohol-mediated cell loss

We examined the role of cGK in preventing alcohol-induced neuronal death in CGN cultures. CGN cultures were pretreated with 8-Br-cGMP, a non-hydrolyzable cGMP analogue that activates cGK. 8-Br-cGMP addition prevented alcohol-induced cell loss (Figs. 1A–B), reducing cell loss from 21.0 ± 2.9% to 4.1 ± 2.3% at the highest dose, 40 μM. Subsequent experiments used an 8-Br-cGMP concentration of 40 μM for two reasons. First, the results in Figs. 1A–B indicate that this concentration significantly reduced alcohol-induced cell loss. Second, 8-Br-cGMP concentrations greater than 40 μM were toxic.

Fig. 1. Cyclic GK activation decreased alcohol-induced neuronal loss.

CGN cultures were pretreated with the cGK-activator, 8-Br-cGMP (six replicates) 30 minutes before alcohol exposure (0 or 400 mg/dl). Cultures in panels C and D also received the cGK inhibitor, Rp-8-pCPT-cGMPS. Left panels (A and C) show cell numbers assessed after 24 hours of alcohol exposure. Alcohol significantly reduced cell numbers (paired t-tests with Bonferroni correction; * p ≤ 0.013). Right panels (B and D) show cell number reductions as percent cell loss. Addition of 8-Br-cGMP at the two highest concentrations provided a dose-dependent neuroprotective effect, reducing the alcohol-induced neuronal loss (repeated measures ANOVA with pairwise comparisons; ^# p ≤ 0.05; paired t-tests, ^& p ≤ 0.025 after Bonferroni correction). The cGK inhibitor, Rp-8-pCPT-cGMPS, eliminated the neuroprotective effect of 8-Br-cGMP.

In addition to activating cGK, 8-Br-cGMP also stimulates cGMP-dependent ion channels [32] and cGMP phosphodiesterases [33], which are both expressed in the cerebellum [32, 33]. To assess whether the neuroprotective effect of 8-Br-cGMP requires cGK, we utilized Rp-8-pCPT-cGMPS, a specific inhibitor of cGK. Results shown in Figs. 1C–D reveal that Rp-8-pCPT-cGMPS (31.3 nM) addition completely eliminated the neuroprotective effect of 8-Br-cGMP, suggesting that the cGMP derivative acts by stimulating the cGMP kinase. Rp-8-pCPT-cGMPS alone produced no neuronal loss.

Cyclic GK activation prevented the alcohol-induced rise in [Ca²⁺]_i

We previously demonstrated that alcohol concentrations of 200, 400, and 800 mg/dl causes a sustained rise in [Ca²⁺]_i linked to cell death in CGN cultures [18]. In the current study, we tested the hypothesis that cGK activation prevents alcohol-induced neuronal death by reducing the associated rise in [Ca²⁺]_i. Thus, we measured [Ca²⁺]_i in Fura-2-loaded CGN cultures receiving cGK-activator, 8-Br-cGMP (0 or 40 μM), prior to alcohol exposure. Since we desired a more robust visual effect in the color rendition of [Ca²⁺]_i, we raised the alcohol concentration to 800 mg/dl only for Fig. 3A (three replicates), but reduced it back to 400 mg/ml for subsequent [Ca²⁺]_i experiments (Figs. 2B–D). Fig. 2A (top row) shows cells before (baseline), and the same cells 60 seconds after alcohol exposure (bottom row). Cells in the top view-fields (Figs. 2Ai–2Aii) are mostly blue, indicating low [Ca²⁺]_i at baseline. In the bottom identical view-fields, more yellow and red cells, indicating high [Ca²⁺]_i, were observed with alcohol treatment (Fig. 2Aiii), than in the group that received 8-Br-cGMP and alcohol (Fig. 2Aiv). Cells in Fig 2Aiv are mostly green and blue indicating low [Ca²⁺]_i. In summary, 8-Br-cGMP greatly reduced the capability of alcohol to raise [Ca²⁺]_i in these cultures.

Fig. 2. Pretreatment with 8-Br-cGMP reduced the alcohol-induced rise in [Ca²⁺]_i.

Panel A shows pseudocolor 340/380 ratio images of CGN cultures (representative for three replicates) before (Ai–Aii) and identical view-fields 60 seconds after adding alcohol (800 mg/dl) (Aiii–Aiv). The color scale bar maps [Ca²⁺]_i; blue and red respectively depict low and high [Ca²⁺]_i. Arrows (Aiii) highlight cells with increased [Ca²⁺]_i after alcohol exposure, shifting in color from blue (low [Ca²⁺]_i) to red/yellow (higher [Ca²⁺]_i). Fewer cells (less arrows) with raised [Ca²⁺]_i are evident in the culture (Aii, Aiv) that received 8-Br-cGMP 30 minutes before alcohol. Panel B: Lines show mean ± SEM [Ca²⁺]_i in cells (combined from six replicates) after adding alcohol (400 mg/dl) over the duration of the experiment (9 minutes). Alcohol initiated a rise in [Ca²⁺]_i which continued for the experiment duration. 8-Br-cGMP pretreatment reduced the alcohol-induced rise of [Ca²⁺]_i. Panel C: Utilizing the data from Panel B, the highest [Ca²⁺]_i attained for each cell (Peak [Ca²⁺]_i) over the duration of the experiment was derived and a mean was calculated. The dashed line indicates the baseline [Ca²⁺]_i level, measured 30 seconds after adding alcohol. Treatment with 8-Br-cGMP decreased the peak [Ca²⁺]_i level elicited by alcohol (* p ≤ 0.05). Panel D: To ascertain the cell distribution across the range of Peak [Ca²⁺]_i, a histogram was derived from the Peak [Ca²⁺]_i data in Panel C. The X-axis (Peak [Ca²⁺]_i) in Fig 3D was divided into 20 nM segments and the percent of cells in each segment was determined. A vertical dashed line is drawn at an [Ca²⁺]_i of 200 nM, and the percentage of cells with more than 200 nM [Ca²⁺]_i was determined. In the absence of 8-Br-cGMP, 27.4% of the cells attained an [Ca²⁺]_i greater than 200 nM, after alcohol addition. Treatment with 8-Br-cGMP substantially lowered this percentage to 4.1% (χ², p ≤ 0.05).

In Figs. 2B–D, five replicate cultures were exposed to alcohol (400 mg/dl). Mean [Ca²⁺]_i was calculated from the [Ca²⁺]_i values of all cells in the CGN cultures and plotted in 5 second intervals (Fig. 2B). Average baseline [Ca²⁺]_i, determined for 30 seconds before alcohol exposure, was 70.3 ± 1.2 nM (882 cells). Following alcohol addition, intracellular [Ca²⁺]_i climbed over the course of experiment (9 minutes). In contrast, the alcohol-induced [Ca²⁺]_i rise was much smaller in 8-Br-cGMP-pretreated cultures. One difference from the 800 mg/dl alcohol-treated cultures (Fig 2A) is that 400 mg/dl produced more gradual [Ca²⁺]_i elevations (Fig 2B), in agreement with our previous study [18]. To evaluate [Ca²⁺]_i responses to alcohol, we compared the maximum “peak” [Ca²⁺]_i attained during the evaluation period (Fig. 2C) [18]. Alcohol alone increased [Ca²⁺]_i, reaching a peak of 163.3 ± 4.8 nM, and never returned to baseline. In comparison, the peak [Ca²⁺]_i elicited by alcohol in cells pretreated with 8-Br-cGMP was substantially lower (117.4 ± 2.4 nM; t-test, p ≤ 0.05). 8-Br-cGMP addition alone did not change basal [Ca²⁺]_i levels (data not shown). Analyzing peak [Ca²⁺]_i, grouped by replicates instead of by individual cells, provided similar results (174.0 ± 25.0 vs. 120.9 ± 12.8 nM; paired t-test, p ≤ 0.05).

Alcohol elicits substantial [Ca²⁺]_i increases for a sizable minority of cells in susceptible CGN cultures [18]. Thus, we tested whether 8-Br-cGMP pretreatment decreased the proportion of cells with large [Ca²⁺]_i responses to alcohol. The histograms in Fig 2D reveal the cellular distribution of [Ca²⁺]_i in response to alcohol exposure (400 mg/dl) and the change in this distribution with 8-Br-cGMP pretreatment (40 μM). Peak [Ca²⁺]_i values for individual cells were grouped into 20 nM increments (Fig. 2D) and the number of cells in each increment was determined [18]. As shown in the top histogram in Fig 2D, alcohol had a robust effect in raising [Ca²⁺]_i to a high level (above 200 nM, highlighted by the vertical line) in a considerable proportion (27.4%) of cells. In contrast, 8-Br-cGMP pretreatment reduced the proportion of cells with [Ca²⁺]_i above 200 nM to 4.1% (bottom histogram in Fig 2D). In summary, 8-Br-cGMP modulates the response of CGN to alcohol by greatly reducing the number of cells reaching high levels of [Ca²⁺]_i (χ², p ≤ 0.05).

Discussion

Alcohol-induced neuronal death has been previously linked to [Ca²⁺]_i disruption. Chelation or preventing Ca²⁺ efflux from intracellular stores is neuroprotective [18-21]. Given the connection between alcohol-induced neuronal death and disruption of [Ca²⁺]_i homeostasis, preventing this disruption is a valid neuroprotective strategy. Our findings here associate NO-cGMP-cGK-mediated neuroprotection to moderating the [Ca²⁺]_i disruption induced by alcohol exposure. Like previous reported [8, 9], cGK activation with 8-Br-cGMP protected CGN cultures from alcohol toxicity, an effect that the cGK inhibitor, Rp-8-pCPT-cGMPS, blocked. Alcohol elevated [Ca²⁺]_i, a key event triggering alcohol-induced neuronal death [18], which cGK activation prevented.

We previously observed that alcohol concentrations of 200, 400, 800 mg/dl produce dose-dependent increases in [Ca²⁺]_i and neuronal death in CGN cultures [18]. Herein, 8-Br-cGMP treatment before alcohol exposure (400 mg/dl) 1) attenuated the rise in mean [Ca²⁺]_i levels; 2) reduced the peak [Ca²⁺]_i that was produced; and 3) decreased the number of cells that attained high [Ca²⁺]_i (≥ 200 nM) levels. These data suggest that cGK activation conferred alcohol resistance by moderating alcohol-induced [Ca²⁺]_i elevations.

Growing evidence connects alcohol-induced neuronal death to [Ca²⁺]_i dysregulation and to endoplasmic reticulum (ER)-related processes [18-21, 34]. IP₃R-dependent ER Ca²⁺ depletion can cause ER stress, excessive mitochondrial Ca²⁺ uptake, and in turn promotes ROS generation and apoptosis [34]. Given this link, preventing disruption of [Ca²⁺]_i homeostasis is a valid neuroprotective strategy.

Cyclic GK phosphorylates multiple substrates that affect [Ca²⁺]_i homeostasis to confer neuroprotective benefits. For instance, alcohol elicits inositol trisphosphate receptor (IP₃R)-mediated release of Ca²⁺ from intracellular stores to trigger cell death in primary neuronal culture [18], rat hippocampal explants [19], and avian embryo models of FASD [21]. IP₃Rs contain serine phosphorylation sites for cGK [24] and multiple isoforms of both proteins are highly expressed in the cerebellum [35, 36]. Yet, both decreased [23, 36, 37] and increased [38, 39] intracellular Ca²⁺ release has been observed after IP₃R phosphorylation. Given these seemingly conflicting literature reports, how cGK affects IP₃R activity likely depends on additional undetermined regulatory factors. One such factor is IP₃R-associated cGK substrate (IRAG), which, when phosphorylated, reduces Ca²⁺ release from all IP₃R isoforms [40]. However, brain IRAG expression is limited to the thalamus [41] and therefore cannot explain the maturation-acquired alcohol resistance that develops in other brain regions. Further investigation is necessary to determine if IP₃Rs play a role in cGK-mediated neuroprotection.

In conclusion, cGK activation via 8-Br-cGMP, prevented both neuronal death and reduced the associated rise in [Ca²⁺]_i from alcohol in CGN cultures. Inhibition of cGK by Rp-8-pCPT-cGMPS prevented cGMP-mediated neuroprotection to alcohol. These findings associate alcohol resistance, mediated by the NO-cGMP-cGK pathway, to mitigating the persistent and toxic alcohol-induced increase in [Ca²⁺]_i. Future studies aimed at the neuroprotective substrates of cGK may provide exciting new insights into alcohol resistance and their connection to alleviating the alcohol-induced disruption of [Ca²⁺]_i homeostasis.

Highlights.

• Cyclic GK activation improved neuronal survival in alcohol-exposed CGN cultures, which a cGK inhibitor blocked.

• Cyclic GK activation reduced both the level of alcohol-induced rise in [Ca²⁺]_i and number of cells that responded to alcohol by increasing [Ca²⁺]_i.