Direct Temporal Analysis of Apoptosis Induction in Living Adherent Neurons

Primary Hippocampal Neuronal Cultures

The hippocampi were obtained from 1-day-old Sprague-Dawley rat pups, following institutional guidelines, and were incubated in dissociation medium (90 mM Na₂SO₄, 30 mM K₂SO₄, 5.8 mM MgCl₂, 0.25 mM CaCl₂, 10 mM kynurenic acid, and 1 mM HEPES with the pH adjusted to 7.4) containing papain (10 U/ml) and cysteine (3 mmol/liter) for two 20-min periods. The hippocampi were then rinsed in dissociation medium and incubated in dissociation medium containing trypsin inhibitor (10-20 U/ml) for three 5-min periods. The neurons were washed in growth medium (Leibovitz's L-15 medium; Gibco BRL, Gaithersburg, MD) containing 6% sterile rat serum (Bioproducts for Science; Indianapolis, IN), 150 mM NaHCO₃, 2.25 mg/ml transferrin, 2.5 μg/ml insulin, 10 nM progesterone, 90 μM putrescine, 15 nM selenium, 35 mM glucose, 1 mM L-glutamine, penicillin, and streptomycin (50 μg/ml), and vitamins. The dissociated neurons were plated at a density of ∼1.5 × 10³ cells/mm² in 35-mm polylysine/laminin-coated plates (Falcon Labware; Lincoln Park, NJ). Neurons were maintained in growth medium at 37C in a humidified atmosphere of 5% CO₂ and 95% room air. All experiments were performed with neurons that had been in culture for 2 weeks. Non-neuronal cells accounted for 10-20% of the total cell population.

Experimental Treatments

NO administration was performed by replacing the culture medium with media containing either sodium nitroprusside (SNP, 300 μM) (Sigma; St Louis, MO) or 6-(2-hydroxy-1-methyl-2-nitrosohydrazino)- N-methyl-1-hexanamine (NOC-9, 300 μM) (Calbiochem; San Diego, CA) for 5 min. We have previously demonstrated that each of these agents yields neuronal injury through a mechanism that involves the direct release of NO, with a 5-min application of 300 μM resuiting in the death of approximately 70-80% of neurons over a 24-hr period (Maiese et al. 1993a,b). More than one NO generator is used as a control to demonstrate that the neurons were responding to NO rather than to other byproducts of these agents. After treatment with the NO donors, the culture medium was replaced with fresh growth medium and the cultures were placed in a normoxic, humidified incubator at 37C with 5% CO₂ for periods determined by the specific experimental protocol.

Staining for Externalization of Phosphatidylserine Residues

Annexin V conjugated to phycoerythrin (PE) was purchased from R&D Systems (Minneapolis, MN). The stock solution was 30 μg/ml concentration. This was diluted directly before use 1:10 in warmed (37C) binding buffer (10 mM HEPES, pH 7.5, 150 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 1.8 mM CaCl₂). The growth medium was removed from culture plates, and annexin V conjugate was applied at a final concentration of 3 μg/ml, and then incubated at 37C in a humidified atmosphere in the dark for 10 min. Plates were then rinsed twice with fresh binding buffer and neurons were examined using a Leitz DMIRB microscope (Leica; McHenry, IL) and Oncor Image 2.0 imaging software (Oncor; Gaithersburg, MD). Images were acquired using a cooled charge-coupled device with both transmitted light and fluorescent single-excitation light at 490 nm and detected emission at 585 nm.

After examination, the annexin V label was detached by washing three times in dissociation buffer (10 mM HEPES, pH 7.5, 150 mM NaCl, 5 mM KCl, 2.8 mM MgCl₂), which differed from binding buffer in that the calcium was replaced with magnesium. Plates could then be reexamined to confirm that the annexin V was completely removed, then returned to the incubator for a further specified period. Plates could then be re-stained using the same method. By drawing a grid on the bottom of the culture dish, the same fields of neurons could be relocated for sequential imaging.

Neuronal Survival Assays

Hippocampal neuronal injury was determined by brightfield microscopy using a 0.4% trypan blue dye exclusion method at specified times after treatment with the NO donors. Neurons were identified by morphology. The mean survival was determined by counting nine randomly selected, nonoverlapping fields with each containing approximately 10-20 neurons (viable + nonviable) in each 35-mm Petri dish. The mean survival from each culture dish represents an n = 1 determination.

Transmission Electron Microscopy of Hippocampal Neurons

Electron microscopy was used to visualize the neuronal subcellular structure to assess changes in the cell nucleus at times relevant to the externalization of phosphatidylserine. Neurons were grown on 35-mm glass coverslips coated with laminin and poly-L-lysine. After treatments as indicated, the hippocampal neurons were fixed at 4C for 1.5 h with a fixative consisting of 1:1:1 of 2% aqueous OsO₄, 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4), and 0.2 M phosphate buffer (pH 7.4). Neurons were then rinsed with phosphate buffer and dehydrated with graded ethanol in the following sequence: 50% for 5 min, 70% for 10 min, 85% for 10 min, 95% for 10 min, 100% for 10 min, and an additional 30 min in 100% ethanol. Neurons were removed from the coverslips and placed in a glass scintillation vial containing propylene oxide (PO). Dehydration was performed with three changes of PO, each for 10 min, and one change for 30 min. Neurons were infiltrated with mixture of 1:1 Epon-Araldite and PO for 1 hr, with 3:1 mixture for 3 hr, with straight Epon-Araldite for 1-4 days, and then with Epon-Araldite-added accelerator for 24-48 hr. Neurons were embedded in fresh Epon-Araldite, and ultrathin sections were mounted on copper grids and stained with 3% aqueous uranyl acetate and Reynolds' lead citrate. These were viewed using a JEM 1010 transmission electron microscope (Jeol; Tokyo, Japan).

Statistical Analysis

For each experiment, the mean and SEM were determined. The sample size is defined in each individual experiment. Statistical significance was assessed using the Student's paired t-test and ANOVA with 95% confidence intervals.

Annexin V Binding to Injured Neurons Is Reversible and Reproducible

Using the reversible annexin V assay, we could determine the time when cultured neurons externalized phosphatidylserine at the plasma membrane. Figure 1 illustrates the reversibility of the annexin V and phosphatidylserine interaction after removal of calcium. Each panel in the figure illustrates the same field of neurons in a culture dish after a representative exposure to the NO generator SNP (300 μM). The neurons were imaged before staining with annexin V-PE (Figures 1A and 1B), during staining with annexin V-PE 5 hr after NO exposure (Figures 1C and 1D), then after rinsing three times in dissociation buffer (Figures 1E and 1F). Three washes with dissociation buffer consistently removed all of the staining in the culture plate. Neurons were then restained 2 hr later with annexin V-PE in the same manner (Figures 1G and 1H).

OPEN IN VIEWER

Figure 1

PCD after free radical injury is identified by reversible annexin V-PE labeling. The figure illustrates a series of images of the same microscope field. (A,C,E,G) Images generated using the transmitted light illustrating the neurons. (B,D,F,H) Corresponding images were obtained using 490-nm excitation and 585-nm emission wavelengths to locate the PE label on the surface of neurons undergoing PCD. Five hours after exposure to SNP (300 μM), the neurons were examined before addition of the label to demonstrate that there was no background fluorescence (B). In C and D, the neurons were stained with annexin V-phycoerythrin conjugate. Stained fluorescent neurons undergoing PCD are shown in D. The label was then removed by washing three times in calcium-free buffer. The efficiency of removal is demonstrated while the neurons were still present (E,F). Neurons could then be re-stained and imaged. (G,H) The neurons that were re-stained 2 hr later demonstrate increased annexin V labeling. Bar = 30 μm.

NO Exposure Induces a Progressive Increase in Phosphatidylserine Residue Translocation

Figure 2 illustrates a series of representative images obtained to characterize changes in annexin V labeling over a 24-hr period. In untreated control cultures, there appeared to be a slight increase in annexin V staining by the 24-hr time point. In neurons exposed to NO, a progressive and significant increase in annexin V labeling was evident over a 24-hr period. Figure 2 illustrates a representative set of images for the NO generator SNP (300 μM). In this series of images, the staining at 3 hr after NO exposure was no different from that at 1 hr. At 5 hr after NO injury, two of the three neurons in this field are annexin V-positive. The intensity of staining in these neurons continued to increase at 7 and 24 hr after NO exposure. The third neuron also became positive for annexin V labeling 7 hr after NO exposure.

OPEN IN VIEWER

Figure 2

Tracking of individual neurons over time illustrates progressive increase in annexin V labeling after free radical injury. Neurons were stained, imaged, and washed repeatedly over 24 hr. Representative fields of stained neurons are shown indicating minimal induction of PCD in control cultures (left column: T, transmitted light; F, fluorescence) and induction of PCD in cultures exposed to SNP (300 μM) (right column: T, transmitted light; F, fluorescence). Over a 24-hr course, a statistically insignificant induction in annexin V labeling of control cultures may be secondary to repeated mechanical manipulation. In contrast to the control cultures, NO exposure resulted in a marked, progressive increase in the labeling of phosphatidylserine residues over a 24-hr course, illustrating the initial onset and progression of PCD. Bars = 20 μm.

We further assessed the degree of annexin V staining by determining the proportion of labeled neurons in each microscope field. Initial percentage of annexin V staining in untreated control cultures at the 1-hr time point was 6 ± 4% (Figure 3). This amount of annexin V staining approached a trend to gradually increase in the same neuronal cells to 17 ± 4% (24 hr) but was not statistically significant (Figure 3). Within 3 hr after the exposure to NO, annexin V labeling is significantly increased from approximately 5 to 20% in the identical neuronal cells (Figure 3), suggesting a rapid induction of PCD. These neurons continue to significantly progress with phosphatidylserine membrane translocation to a maximum of approximately 60% over a 24-hr time period (Figure 3). Therefore, our ability to follow the temporal course of phosphatidylserine membrane translocation in individual neurons after free radical exposure with NO demonstrates that the resultant induction of PCD is both an early (within 3 hr) and a robust process that serially progresses in individual neurons over a 24-hr course.

OPEN IN VIEWER

Figure 3

Quantitation of neurons undergoing PCD. At each indicated time point after exposure to either SNP (300 μM) or NOC-9 (300 μM), the percentage of neurons labeled with annexin V was determined. By 3 hr, a significant increase in annexin V label was present after NO exposure. Over a 24-hr period, annexin V label progressively increased to a maximum of approximately 60%. Data represent the mean and SEM from seven individual experimental cultures. In each experiment, the percentage of stained neurons was counted in three to seven discrete fields with 5-25 neurons in each field. Results of neuronal cultures exposed to NO (^∗) were significantly different from control (untreated) cultures not exposed to NO. p<0.05 by ANOVA.

Annexin V Binding Identifies Neurons Early in PCD Before the Loss of Membrane Integrity

Translocation of membrane phosphatidylserine residues has been reported to be an early marker of PCD induction (Rimon et al. 1997). If this is the case, then positive labeling of externalized phosphatydylserine residues should occur before the onset of cell membrane disruption. We therefore characterized the ability of annexin V to identify early PCD induction before cell membrane leakage with a dye exclusion system in sister cultures (Figure 4). The loss of membrane integrity was assessed by counting the proportion of trypan blue-positive neurons as described for the survival assays. In this experiment, the NO generators NOC-9 (300 μM) and SNP (300 μM) were employed. To simplify the analysis, results for the two NO donors were combined. Loss of membrane integrity increased by insignificant increments from approximately 2% at 1 hr to 12% at 7 hr after NO exposure (Figure 4). After the 7-hr time period, trypan blue staining markedly increased to 72 ± 2%. In contrast, annexin V labeling rapidly and significantly outpaced the induction of cellular membrane disruption at the time points of 3, 5, and 7 hr after NO exposure (Figure 4). Our work illustrates that annexin V binding identifies neuronal injury and the dislruption of PCD before the loss of cellular membrane integrity.

OPEN IN VIEWER

Figure 4

Phosphatidylserine exposure precedes the loss of membrane integrity. At each indicated time point after exposure to either SNP (300 μM) or NOC-9 (300 μM), the percentage of neurons labeled with annexin V was determined. In sister cultures, the staining with trypan blue was counted at the same times after identical NO exposure. To simplify the figure, the results for the two NO donors were combined. In each experiment, the percentage of annexin V-labeled neurons was counted in three to seven discrete fields with 5-25 neurons in each field. The trypan blue data represent the mean and SEM from two separate experimental cultures, counting nine randomly selected, nonoverlapping fields with each containing approximately 10-20 neurons in each culture dish. ^∗ The annexin V labeling at 3, 5, and 7 hr after NO exposure was significantly higher than the induction of cellular membrane leakage of trypan blue (p<0.001, ANOVA).

As shown in Figure 5, we further assessed the ability of neurons labeled with annexin V to simultaneously stain with trypan blue. After identification of neurons that label for annexin V at each time period, warmed (37C) 0.4% trypan blue was added in a 1:1 volume:volume dilution for 1 min, removed, and then neurons were re-imaged. At the time points of 3 hr, 5 hr, and 7 hr, the majority of neurons labeling positive for annexin V did not stain for trypan blue. Figure 5 is a representative image at the 5-hr time point, which illustrates that annexin V labeling of phosphatidylserine residue translocation precedes cell membrane disruption that is detected with trypan blue staining. In contrast, over a 24-hr period, labeling with annexin V coincides with trypan blue staining in the same neuronal population (Figure 4), suggesting that PCD has progressed to a level that now involves cell membrane disruption. These results correlate with our prior studies that demonstrated PCD expression and trypan blue staining in 70% of primary hippocampal neurons 24 hr after NO exposure (Maiese et al. 1993a; Maiese and Boccone 1995; Vincent et al. 1997). Our studies also illustrate that some neurons that lose membrane integrity, as evidenced by trypan blue staining, no longer label with annexin V (Figure 5). These results are consistent with the findings in other cellular systems (Rimon et al. 1997) that loss of annexin V labeling after severe membrane disruption may be secondary to shedding of phosphatidylserine residues or the loss of membrane fluidity. Therefore, our method is sensitive for assessment of the initial stages of PCD induction before more “global” cellular injury.

OPEN IN VIEWER

Figure 5

Annexin V-labeled neurons do not stain with trypan blue during early cell injury. The images are representative and were obtained from the same field of neurons 5 hr after NOC-9 (300 μM) exposure. The left panel (T) illustrates the transmitted light image after addition of trypan blue dye. The black arrow identifies a neuron unable to exclude trypan blue. In the corresponding fluorescent image (F), the annexin V-labeled neurons are indicated with white arrows and the neuron unable to exclude trypan blue is identified by the black arrow. The majority of neurons, including the annexin V-labeled neurons, did not take up the trypan blue dye, indicating that their cell membranes were intact. Bar = 20 μm.

Annexin V Staining Does Not Independently Alter Neuronal Survival

Depending on the mode and concentration of application, annexin V application can have either toxic or neurotrophic effects on neurons (Takei et al. 1994). Neurotrophic effects have been reported with a continuous 5-day exposure within a low concentration range (1-30 ng/ml). Although we employ a brief application of annexin V for 10 min, our concentration of annexin V (3 μg/ml) is greater than concentrations that have been demonstrated to yield neurotrophic effects. For these reasons, we investigated the ability of annexin V to alter neuronal survival in our primary hippocampal cultures. Assessing neuronal survival over a 24-hr period, repeated application of annexin V at 1 hr, 3 hr, 5 hr, 7 hr, and 24 hr slightly decreased neuronal survival from 72 ± 2% (untreated control) to 61 ± 2% (repeated annexin application) (Figure 6). However, application of annexin V in conjunction with an NO generator did not significantly alter neuronal survival compared to application of an NO generator alone. In light of these results, annexin V application appears not to be directly toxic to the neurons, but a 10% decrease in neuronal survival may be secondary to mechanical cellular disruption after repeated washing and staining applications over a 24-hr period. The neuronal injury that follows this mechanical damage may occur through necrosis or PCD.

OPEN IN VIEWER

Figure 6

Repeated staining with annexin V does not alter neuronal survival. Neuronal survival was assessed over a 24-hr period after repeated application of annexin V at 1 hr, 3 hr, 5 hr, 7 hr, and 24 hr. Although neuronal survival slightly decreased from 72 ± 2% (untreated control) to 61 ± 2% [annexin V only; p<0.05 by ANOVA (^∗)], application of annexin V in conjunction with an NO generator did not significantly alter neuronal survival compared to application of an NO generator alone. The results demonstrate that annexin V application is not directly toxic to the neurons, but a 10% decrease in neuronal survival may be secondary to mechanical cellular disruption after repeated washing and staining applications over a 24-hr period. Data represent the mean and SEM from two separate experimental cultures.

Annexin V Staining Correlates Closely with Changes in Nuclear Morphology

Our present results illustrate that our technique offers a sensitive approach for the early detection of the induction of PCD. However, it also is vital to assess the specificity of our assay to detect PCD. Electron microscopy provides a secondary means of visualizing changes in the neuronal subcellular structure that are consistent with PCD. Primary hippocampal neurons were exposed to an NO generator and then processed 1 hr, 3 hr, 5 hr, 7 hr, and 24 hr later for transmission electron microscopy. Evidence for apoptotic neuronal cell death was characterized by the preservation of membrane integrity and internal organelle structure and by the presence of chromatin condensation with nuclear fragmentation (Figure 7). Assessment by electron microscopy for PCD closely paralleled PCD characterization through the annexin V labeling. Approximately 10% or fewer neurons displayed evidence of PCD after 1 hr, with the majority of neurons showing no evidence of PCD (Figure 7A). A slightly greater proportion (10-20%) of neurons displayed some chromatin condensation after 3 hr (Figure 7B). At 5 hr, approximately 35% of the neurons were identified with chromatin condensation (Figure 7C). The percentage of neurons on electron microscopy that were consistent with PCD increased to approximately 50% by 7 hr (Figure 7D) and to almost 70% by 24 hr (not shown).

OPEN IN VIEWER

Figure 7

Nuclear morphological changes correlate with annexin V labeling. Transmission electron microscopy was employed to examine subcellular manifestations of PCD after NO exposure. Each image is representative of a neuronal nucleus at 1 hr (A), 3 hr (B), 5 hr (C), and 7 hr (D) after exposure to NOC-9 (300 μM). The figure demonstrates the progressive appearance of condensed chromatin with margination at the nuclear membrane. Nuclear cleavage began to appear at 5 hr after NO (C) and became more evident at 7 hr after NO (D). Bar = 2 μm.