Epilepsy is characterized by the occurrence of spontaneous recurrent epileptiform discharges (SREDs) in neurons. A decrease in calcium/calmodulin-dependent protein kinase II (CaMK-II) activity has been shown to occur with the development of SREDs in a hippocampal neuronal culture model of acquired epilepsy, and altered calcium (Ca2+) homeostasis has been implicated in the development of SREDs. Using antisense oligonucleotides, this study was conducted to determine whether selective suppression of CaMK-II activity, with subsequent induction of SREDs, was associated with altered Ca2+ homeostasis in hippocampal neurons in culture. Antisense knockdown resulted in the development of SREDs and a decrease in both immunocytochemical staining and enzyme activity of CaMK-II. Evaluation of [Ca2+]i using Fura indicators revealed that antisense-treated neurons manifested increased basal [Ca2+]i, whereas missense-treated neurons showed no change in basal [Ca2+]i. Antisense suppression of CaMK-II was also associated with an inability of neurons to restore a Ca2+ load. Upon removal of oligonucleotide treatment, CaMK-II suppression and Ca2+ homeostasis recovered to control levels and SREDs were abolished. To our knowledge, the results demonstrate the first evidence that selective suppression of CaMK-II activity results in alterations in Ca2+ homeostasis and the development of SREDs in hippocampal neurons and suggest that CaMK-II suppression may be causing epileptogenesis by altering Ca2+ homeostatic mechanisms.
Epilepsy, a common neurological disorder affecting approximately 1 to 2% of the population worldwide (Hauser and Hesdorffer, 1990; McNamara, 1999), is characterized by the occurrence of spontaneous recurrent epileptiform discharges (SREDs) in populations of neurons. Acquired epilepsy (AE) is caused by a known brain injury, such as status epilepticus (SE), stroke, or traumatic brain injury that induces longlasting plasticity changes in previously normal brain tissue and leads to the development of SREDs (Hauser and Hesdorffer, 1990). Epileptogenesis is the process responsible for transforming normal brain tissue into the epileptic phenotype, manifesting the occurrence of seizures or SREDs (Shin and McNamara, 1994; DeLorenzo et al., 1998, 2005). Although considerable research is being conducted on longterm changes associated with epileptogenesis, the exact cellular and molecular mechanisms involved in epileptogenesis are still not completely understood (Shin and McNamara, 1994; Delorenzo et al., 2005). Therefore, studies elucidating the underlying mechanisms mediating the plasticity changes associated with epileptogenesis are crucial to advancing both the treatment and prevention of epilepsy.
It has been suggested that chronic elevations in intracellular Ca2+ ([Ca2+]i) and alterations in Ca2+ homeostatic mechanisms in epileptic neurons play a role in the induction and maintenance of the epileptic phenotype (DeLorenzo et al., 1998, 2005; Raza et al., 2001, 2004). Calcium is an important second messenger involved in diverse cellular events, such as membrane excitability and synaptic activity (Delorenzo et al., 2005). Elevated [Ca2+]i can lead to excitotoxicity, which precedes epileptogenesis and epilepsy (DeLorenzo et al., 1998, 2005; Raza et al., 2004; Sun et al., 2004). In normal cells, [Ca2+]i is regulated by several Ca2+ homeostatic mechanisms (Carafoli et al., 1997). Studies in epileptic neurons have shown that disruption of Ca2+ extrusion and sequestration mechanisms can lead to elevated basal [Ca2+]i with a decreased ability to restore resting [Ca2+]i levels following an external Ca2+ load (Pal et al., 2001; Raza et al., 2001; Sun et al., 2004). Although many studies have linked increases in [Ca2+]i and altered Ca2+ homeostatic mechanisms to epilepsy (Delorenzo et al., 2005), it remains unclear what molecular events during epileptogenesis cause these changes in Ca2+ homeostasis.
Calcium-calmodulin-dependent protein kinase II (CaMK-II) plays a major role in modulating neuronal excitability and function (Kelly et al., 1984), with alterations in CaMK-II levels linked to neuronal hyperexcitability. Decreases in CaM kinase II have been reported in numerous in vivo and in vitro models of epilepsy (Bronstein et al., 1993), including kindling (Wasterlain and Farber, 1984; Taft et al., 1987), electrical stimulation SE (Perlin et al., 1992), pilocarpine (Churn et al., 2000a), and low Mg2+ in cultured neurons (Blair et al., 1999). Furthermore, CaMK-II knockout mice demonstrated the epileptic phenotype and developed spontaneous seizures (Butler et al., 1995). It has been demonstrated that knocking down CaMK-II activity in cultured hippocampal neurons with an antisense oligonucleotide resulted in epileptiform activity as evidenced by the presence of SREDs (Churn et al., 2000b). It was also shown in cultured hippocampal neurons that CaMK-II activity was decreased in association with the development of SREDs (Blair et al., 1999). These studies implicate a role for altered CaMK-II function toward the induction and maintenance of the epileptic phenotype and suggest that alterations in CaMK-II activity may play a role in the induction of epileptogenesis by altering Ca2+ homeostatic mechanisms.
This study was initiated to evaluate whether the decrease in CaMK-II activity during epileptogenesis could represent a molecular mechanism that underlies alterations in Ca2+ homeostasis observed in the epileptic phenotype. Antisense oligonucleotides were used to experimentally suppress CaMK-II function in hippocampal neurons in culture. Following knockdown of CaMK-II, electrophysiological studies were initiated to determine the presence or absence of SREDs, and Ca2+-imaging studies were employed using the fluorescent calcium indicators Fura-2 and Fura-FF to evaluate [Ca2+]i dynamics. Similar experiments were performed on neurons that were allowed to recover from antisense knockdown of CaMK-II. Antisense oligonucleotide suppression of α-CaMK-II in neurons resulted in the development of SREDs, higher basal [Ca2+]i, and a decreased ability to restore resting [Ca2+]i after a Ca2+ load compared with control and antisense-recovered neurons. The results provide the first direct evidence that CaMK-II activity is involved in maintaining Ca2+ homeostasis in neurons and that the alterations in Ca2+ homeostasis resulting from suppression of CaMK-II may underlie the development of SREDs. Because of the pathophysiological similarities to the glutamate injury and low Mg2+ in vitro models of AE, knocking down CaMK-II levels in cultured hippocampal neurons provide a novel model for studying epileptiform activity.
Materials and Methods
Hippocampal Neuronal Culture. Primary hippocampal cultures were prepared as described previously by our laboratory (Sombati and Delorenzo, 1995). In brief, hippocampal cells were isolated from 2-day postnatal Sprague-Dawley rats (Harlan, Indianapolis, IN) and plated at a density of 2.4 × 104 cells/cm2 onto glial support layers previously plated onto poly-l-lysine (0.05 mg/ml)-coated Lab-Tek two-well cover glass chambers (Nalge-Nunc International, Naperville, IL). Cultures were maintained at 37°C in a 5% CO2/95% O2 air atmosphere and fed twice weekly with Neurabasal A medium (Invitrogen, Carlsbad, CA) enriched with B-27 serum-free supplement (Invitrogen) and 0.5 mM l-glutamine. Cultures were exposed to 5 μM cytosine arabinoside 2 days after plating to inhibit non-neuronal growth. Experimental studies and [Ca2+]i measurements were performed on neurons maintained for 15 to 18 days in vitro (DIV) to ensure adequate neuronal development.
αCaMK-II Antisense Knockdown Analysis. Antisense and missense oligonucleotide probes (Operon Biotechnologies, Huntsville, AL) were constructed as described previously (Churn et al., 2000b). For the α-CaMK-II knockdown, an antisense oligonucleotide complimentary to the +1 to +18 nucleotides was constructed with the sequence 5′-GGT AGC CAT CCT GGC ACT-3′; the missense sequence for α-CaMK-II was 5′-GGT AGC CAT AAG GGC ACT-3′. Knockdown treatment for αCaMK-II involved treating hippocampal cultures for 3 days with a 3.0 μM concentration of either antisense or missense oligonucleotides every 24 h, as determined by measuring CaMK-II activity (Churn et al., 2000b). Oligonucleotide treatment was initiated at 13 DIV to allow for analysis to occur by 17 DIV. Untreated neurons were used as controls. During the treatment regimen, hippocampal cultures were maintained in Neurobasal A medium supplemented with B-27. At the end of the 3-day treatment protocol, cultures were utilized for Fura-2 [5-oxazolecarboxylic acid, 2-(6-(bis(2-((acetyloxy)methoxy)-2-oxoethyl)amino)-5-(2-(bis(2-((acetoxy(methoxy)-2-oxoethyl)amino-5-methylphenoxy)ethoxy)-2-benzofuanyl)-,(acetoxy)methyl ester] and Fura-FF [Ca2+]i measurements. To assess the recovery and restoration of CaMK-II activity, cultures were treated with antisense or missense oligonucleotide every 24 h for 3 days. After the 3rd day of treatment, oligonucleotide treatments were discontinued and the neurons were allowed to recover in oligonucleotide-free medium. After full recovery of CaMK-II activity, cultures were utilized for [Ca2+]i measurements.
Measurement of CaMK-II Activity. CaMK-II activity was determined measuring CaMK-II-dependent substrate phosphorylation of the synthetic peptide Syntide II (Sigma, St. Louis, MO) following the methods of Churn (Churn, 1995). Phosphorylation reaction solutions contained 41 μg of protein, 60 μM Syntide II, 10 mM MgCl2, 7 μM [γ-32P]ATP, 10 mM PIPES, pH 7.4, ± 5 μM CaCl2, and ± 1 μg of calmodulin. Standard reactions were performed in a shaking water bath at 30°C. The phosphorylation reactions were initiated by adding Ca2+, continued for 1 min, and stopped by adding 20 μM EDTA. Ten-microliter aliquots of assay solution were blotted onto P-81 phosphocellulose filter paper (Whatman Inc., Florham Park, NJ). Each reaction was quantitated in triplicate. The filter paper was then washed three times in 50 mM phosphoric acid, rinsed with acetone, and allowed to air-dry. Radioactive phosphate was then quantitated by scintillation counting (Churn, 1995).
Immunocytochemical Staining of α-CaMK-II. After briefly washing in phosphate-buffered saline (PBS), neuronal cultures were fixed in 4% paraformaldehyde in PBS for 10 min followed by a 3 × 5-min wash in PBS. Fixed cultures were blocked and permeabilized in SuperBlock blocking buffer (Pierce, Rockford, IL) containing 0.2% Triton X-100 for 60 min at room temperature. Cells were then incubated with a mouse monoclonal antibody to the CaMK-II α-subunit (10 μg/ml, clone 6G9; Biomol, Plymouth Meeting, PA) in SuperBlock blocking buffer containing 0.1% Triton X-100 overnight at 4°C. Cells were washed 4 × 5 min in PBS containing 0.1% Triton X-100. After the wash, cells were incubated with Texas Red (Vector Laboratories, Burlingame, CA) conjugated anti-mouse IgG (20 μg/ml) in Super-Block blocking buffer for 60 min at room temperature. Cells were washed 4 × 5 min in PBS containing 0.1% Triton X-100 followed by one wash in PBS alone. Labeled cells were coated with Vectashield (Vector Laboratories) and coverslipped. Control staining was carried out in an identical manner, with the exception of removal of the primary antibody. Fluorescence microscopy was carried out on an Olympus inverted microscope (Olympus America) fitted with a 20× objective and a Texas Red filter cube allowing for excitation/emission of 595/615 nm. Fluorescent images were captured with a Q-fire digital camera and evaluated with Olympus MicroSuite imaging software (Soft Imaging System, Lakewood, CO).
Whole-Cell Current-Clamp Analysis of Hippocampal Neuronal Cultures. Electrophysiological analysis was performed using previously established procedures in our laboratory (Sombati and Delorenzo, 1995). In brief, cell culture media were replaced with physiological bath recording solution (pBRS) containing 145 mM NaCl, 2.5 mM KCl, 10 mM HEPES, 10 mM glucose, 2 mM CaCl2, 1 mM MgCl2, and 2 μM glycine, pH 7.3; osmolarity was adjusted to 325 mOsm with sucrose. Cell culture plates were mounted on the stage of an inverted microscope (Nikon Diaphot, Japan), continuously perfused with pBRS, and then studied using the whole-cell currentclamp recording procedure. Patch electrodes with a resistance of 2 to 4 MΩ were pulled on a Brown-Flaming P-80C electrode puller (Sutter Instruments, Novato, CA) and then fire-polished. For whole-cell current-clamp analysis, the electrode was filled with a solution containing 140 mM K+ gluconate, 1 mM MgCl2, and 10 mM Na-HEPES, pH 7.2; osmolarity was adjusted to 310 ± 5 mOsm with sucrose. Data were digitized and transferred to videotape using a pulse code modulator device (Neurodata, New York, NY) and then played back on a DC-500 Hz chart recorder (Astro-Med Dash II, Warwick, RI). Intracellular recordings were carried out using an Axopatch 200B amplifier (Axon Instruments, Foster City, CA) in whole-cell current-clamp mode.
Intracellular Free Calcium Measurements. Intracellular free Ca2+ measurements were carried out using previously established procedures (Pal et al., 1999, 2001; Raza et al., 2001; Sun et al., 2004). Hippocampal neurons were loaded with acetoxymethyl form of the membrane-permeable ratiometric fluorescent Ca2+ indicators Fura-2 or Fura-FF (1 μM; Invitrogen) in pBRS for 30 min at 37°C. Loaded cells were washed three times with pBRS and incubated for an additional 15 min to allow for cellular esterase cleavage of the acetoxymethyl moiety and intracellular trapping of the free acid Fura-2 or Fura-FF indicators. Culture dishes were mounted onto a heated stage at 37°C on an Olympus IX-70 inverted microscope fitted with a 20× fluorite water immersion objective and coupled to an ultra high-speed fluorescence imaging system (Olympus America/PerkinElmer Life and Analytical Sciences, Boston, MA). The fluorescence excitation source was a 75-W Xenon arc lamp (Olympus America). Neutral density filters of variable opacities were used to attenuate unwanted excitation. Alternating excitation wavelengths of 340 and 380 nm were generated using a Lambda 10-2 filter wheel (Sutter Instruments Co., Novato CA), and 510-nm emissions were acquired using a Fura filter cube (Olympus America) with a dichroic at 400 nm. Alternating emissions from 340- and 380-nm excitation were captured with an ORCA-ER high-speed digital CCD camera (Hamamatsu Photonics K.K., Hamamatsu City, Japan). Image acquisition and processing were computer-controlled using the Ultra-VIEW Imaging system software version 5.2 (PerkinElmer Elmer Life and Analytical Sciences). Using the image analysis software, a region of interest was selected for each pyramidal-shaped neuron in the field, and intracellular free Ca2+ levels were presented as an absolute ratio value of 340/380-nm excitation-emissions. To calculate 340/380 ratio values, image pairs at each wavelength were captured and digitized at varying intervals, and the images at each wave-length were averaged over four frames.
Background autofluorescence values for both 340- and 380-nm excitations were obtained by imaging a field of neurons lacking the fluorescent indicator and were subtracted from experimental 340/380-nm excitation-emissions values from indicator-loaded cells. Absolute ratio values of 340/380-nm excitation-emissions of intracellular free Ca2+ levels were used in this study as a quantitative measure of [Ca2+]i levels, employing established techniques (Pal et al., 1999, 2000; Raza et al., 2001; Sun et al., 2004).
Experimental studies were carried out on oligonucleotide-treated hippocampal cultures to evaluate [Ca2+]i dynamics. In some experiments, recording solution during Ca2+ imaging contained 400 nM tetrodotoxin (TTX; Sigma) to block synaptic transmission to allow for measurement of isolated [Ca2+]i dynamics in the absence of seizure and synaptic activity. Multiple neuronal culture fields were evaluated to determine basal [Ca2+]i measurements, each for 60 s with image/ratio values acquired at 5-s intervals. For evaluation of the glutamate-induced Ca2+ loads, basal [Ca2+]i levels were recorded for 3 min at 30-s intervals before glutamate treatment. The recording solution was replaced with 1 ml of pBRS containing glutamate (50 μM; Sigma), and glycine (10 μM) was added to the cells. After 2 min of exposure, the glutamate was removed, and the cells were washed twice in pBRS. The cells were monitored for up to 30-min post-glutamate treatment at 30-s intervals.
Calcium Calibration. The [Ca2+]i values in the paper are presented as 340/380 ratio values for Fura-2 and Fura-FF, because absolute [Ca2+]i concentrations can vary depending on the fluorescent indicator used (Hyrc et al., 1997; Pal et al., 2000; Raza et al., 2001). It has been reported that ratio values provide more objective presentation of [Ca2+]i (Pal et al., 2000; Raza et al., 2001). However, we have performed Ca2+ calibration determinations on both the Fura-2 and Fura-FF data to provide estimates of absolute [Ca2+]i concentrations from the 340/380 ratio values using established procedures (Pal et al., 2001). An in vitro calcium calibration curve was constructed for each indicator and used to convert fluorescent ratios to [Ca2+]i concentrations using ionomycin (10 μM) and prepared calibration buffers (Invitrogen, Carlsbad, CA). [Ca2+]i concentrations were calculated from the background corrected 340/380 ratios using the following equation (Grynkiewicz et al., 1985): where R is the 340/380 ratio at any time; Rmax is the maximum measured ratio in saturating Ca2+ solution (39 μM free Ca2+); Rmin is the minimal measured ratio Ca2+ free solution; Sf2 is the absolute value of the corrected 380-nm signal at Rmin; Sb2 is the absolute value of the corrected 380-nm signal at Rmax; and the Kd values used were 224 nM and 20 μM for Fura-2 and Fura-FF, respectively (Pal et al., 2001; Sun et al., 2004).
For basal [Ca2+]i experiments, Fura-2 was used as the indicator, because Fura-FF is not sensitive in the low nanomolar range (Hyrc et al., 1997). For the experiments evaluating Ca2+ homeostasis, both Fura-2 and Fura-FF were used because of the large change in [Ca2+]i produced by the glutamate exposure. Basal [Ca2+]i concentrations for control, missense-treated, antisense-treated, and antisense-recovered neurons were determined from the ratio values and were 107.1 ± 3.4, 95.1 ± 3.4, 138.4 ± 7.9, and 110.3 ± 4.6 nM, respectively. For the calcium homeostasis experiments, peak glutamateinduced [Ca2+]i values in missense-treated and antisense-treated neurons using Fura-2 were calculated to be 1.99 ± 0.16 and 1.88 ± 0.36 μM, respectively. Peak glutamate-induced [Ca2+]i concentrations in missense-treated and antisense-treated neurons employing Fura-FF were determined to be 29.3 ± 1.4 and 33.8 ± 2.7 μM. The values obtained in this study for peak glutamate-induced [Ca2+]i values using Fura-2 and Fura-FF were essentially identical to values obtained previously in both in vitro and in vivo epileptic neurons (Pal et al., 2000; Raza et al., 2001). Although the peak [Ca2+]i concentration varied between Fura-2 and Fura-FF, as previously observed (Pal et al., 2000), the ability of the indicators to reflect the neurons' Ca2+-buffering capability was similar. Although Fura-2 gave a lower absolute calcium concentration upon calibration, the shapes of its decay curves were almost identical to those of Fura-FF.
Data Analysis. [Ca2+]i data were collected using the UltraVIEW Imaging System and statistically analyzed and plotted using SigmaPlot (version 8.0) software. Experiments were repeated at least three times from different batches of neuronal cultures. The significance of the data were tested by Student's t test, one-way ANOVA, or one-way repeated measures (RM) ANOVA where applicable The Tukey's test was used as the post hoc analysis for multiple comparisons. Statistical analysis was performed using SigmaStat 2.0 (Jandel Corp., San Rafael, CA). P < 0.05 was considered statistically significant for all data analysis. To compare the kinetics of the [Ca2+]i decay curves, ratio values from individual neurons obtained after glutamate exposure were normalized to “Percent of [Ca2+]i Load” using the following equation: where Ratiot = x is the 340/380 ratio at any time following the glutamate exposure; Ratiot = 0 is the 340/380 ratio at the peak [Ca2+]i load; and RatioBasal is the average 340/380 ratio before glutamate exposure.
Suppression of CaMK-II Activity and Protein Expression in Hippocampal Neurons in Culture. These studies were initiated to evaluate whether the decreased activity and expression of CaMK-II caused by antisense oligonucleotide suppression specific for the α subunit of CaMK-II were reversible. Hippocampal neurons in culture were exposed to either missense or antisense oligonucleotides specific for the α subunit of CaMK-II. CaMK-II-dependent substrate phosphorylation of the synthetic peptide Syntide II was carried out to evaluate the activity of CaMK-II following oligonucleotide treatment (Fig. 1A). Immunocytochemical staining was performed to determine the localization of the α subunit of CaMK-II in cultures treated with missense or antisense oligonucleotides (Fig. 1B). Antisense oligonucleotide treatment significantly decreased CaMK-II activity and protein expression in comparison to control and missense-treated neurons (Fig. 1, A and B). Removal of the antisense oligonucleotide treatment followed by a 3-day recovery period resulted in a restoration of CaMK-II activity and protein expression to missense control levels (Fig. 1, A and B, bottom), demonstrating that the antisense oligonucleotide suppression of CaMK-II was reversible.
Suppression of CaMK-II Activity Correlates with the Development of Epileptiform Activity. Whole-cell current-clamp recordings were performed on pyramidal neurons from control, missense-treated, antisense-treated, and antisense-recovered cultures (Fig. 2). Recordings from missense-treated cultures showed the absence of SREDs and manifested similar intrinsic baseline activity as seen in control cultures (Fig. 2, A and B). Antisense oligonucleotide suppression of CaMK-II resulted in the expression of SREDs in hippocampal neurons (Fig. 2C). Expansion of one of the SREDs from the recording in Fig. 2C revealed the presence of paroxysmal depolarizing shifts with high-frequency spiking activity, characteristic of epileptiform discharges (Fig. 2Ca). The presence of intermittent SREDs remained constant during the antisense oligonucleotide suppression of CaMK-II. Following withdrawal of antisense oligonucleotide treatment with subsequent recovery of CaMK-II activity and expression (Fig. 1, A and B), SREDs were no longer observed, and neuronal electrophysiological activity returned to control levels (Fig. 2D). These results demonstrate that recovery from antisense-oligonucleotide treatment restored CaMK-II activity and completely abolished SREDs, strongly indicating a causal relationship between CaMK-II suppression and the development of SREDs.
Suppression of CaMK-II Activity Results in Elevation of Basal [Ca2+]i. Since alteration of [Ca2+]i has been implicated in epileptogenesis (Delorenzo et al., 2005), we wanted to investigate whether the suppression of CaMK-II, with subsequent induction of SREDs, affected neuronal [Ca2+]i. To evaluate the effects CaMK-II knockdown on [Ca2+]i in neurons, [Ca2+]i imaging was conducted on control, missense-treated, antisense-treated, and antisense-recovered neurons using the fluorescent Ca2+ indicator Fura-2. Figure 3 shows basal [Ca2+]i measurements from control (n = 124), missense-treated (n = 164), and antisense-treated (n = 164) neurons. Treatment with missense oligonucleotide did not alter basal [Ca2+]i, with average 340/380 ratios from control and missense-treated neurons at 0.26 and 0.25, respectively. Neurons from antisense-treated cultures had significantly higher basal [Ca2+]i compared with both control and missense-treated neurons, with an average ratio value of 0.31 (p < 0.001, ANOVA with Tukey's post hoc test), demonstrating that treatment with antisense oligonucleotide with subsequent suppression of CaMK-II function caused an increase in basal neuronal [Ca2+]i (Fig. 3A).
To determine whether the effect of CaMK-II knockdown on basal [Ca2+]i levels was reversible, treatment with antisense oligonucleotide to αCaMK-II was discontinued and neurons were allowed 3 days to recover. As previously shown, SREDs were no longer observed, and the activity and protein expression of CaMK-II were restored to normal levels following recovery (Figs. 1 and 2). Using Fura-2, basal [Ca2+]i was measured in CaMK-II antisense-recovered neurons (n = 86). After the recovery period, ratio values had significantly declined to 0.27 (Fig. 3, P < 0.001 ANOVA with Tukey's post hoc test). These results indicate that the alterations in basal [Ca2+]i induced by the antisense oligonucleotide suppression of CaMK-II were reversible upon a 3-day recovery period following removal of antisense treatment.
Figure 4 shows a histogram breaking down the percentage of neurons from each treatment group into specific ratio value ranges. The distribution of basal [Ca2+]i in neurons from this study are similar to the distribution of basal [Ca2+]i in neurons from the stroke-induced epilepsy model (Sun et al., 2004). A greater percentage of control (Fig. 4A) and missense-treated neurons (Fig. 4B) demonstrated lower basal [Ca2+]i, with 80.8% of control neurons and 78.1% of missense-treated neurons having a ratio value of 0.3 or less. A greater percentage of antisense-treated neurons (Fig. 4C) demonstrated higher basal [Ca2+]i. Only 51.0% of neurons had a ratio value of 0.3 or less. A greater percentage of neurons from antisense-recovered cultures returned to the lower ratio ranges compared with nonrecovered neurons (Fig. 4D), as demonstrated by a leftward shift of the histogram; 81.9% of the antisense-recovered neurons had ratio values of 0.3 or less.
Alteration of Ca2+Homeostasis in Neurons with Decreased Expression of αCaMK-II. To evaluate the ability of the neurons to regulate [Ca2+]i homeostasis, a Ca2+ load was induced by brief exposure to glutamate resulting in an immediate and marked increase in [Ca2+]i using established techniques (Pal et al., 2001; Sun et al., 2004). Both Fura-2 and Fura-FF were employed to evaluate changes in [Ca2+]i caused by glutamate exposure. [Ca2+]i measurements were taken from missense-treated and antisense-treated neuronal cultures. Following brief glutamate exposure (50 μM, 2 min), the glutamate-induced Ca2+ load and the ability of each treatment group to regulate [Ca2+]i was analyzed. Analysis of the glutamate-induced [Ca2+]i load in missense-treated and antisense-treated neurons indicated that the increased [Ca2+]i during the glutamate exposure was not significantly different in both Fura-2 and Fura-FF-loaded cultures. The change in the 340/380 ratios before and during glutamate in cultures loaded with Fura-2 was 1.07 ± 0.05 for missense-treated cultures and 0.93 ± 0.09 for antisense-treated cultures (Fig. 5A; P = 0.1, Student's t test). In Fura-FF-loaded cultures, the change in the 340/380 ratios before and during glutamate was 0.19 ± 0.01 for missense-treated cultures and 0.20 ± 0.02 for antisense-treated cultures (Fig. 5B; P = 0.8, Student's t test). Thus the impairment of the antisense-treated neurons to buffer [Ca2+]i was not due to an increased [Ca2+]i load produced by glutamate.
Because the increase in 340/380 ratios during glutamate exposure was not different between missense-treated and antisense-treated neurons, we normalized the post-glutamate ratio values of individual neurons to the ratio values of their Ca2+ loads to compare Ca2+ decay curves between missense-treated and antisense-treated neurons (Fig. 5, C and D). Individual curves, normalized as percentage of Ca2+ load, were averaged and compared using the RM ANOVA (Sun et al., 2004). By 30 min post-glutamate, missense-treated neurons recovered to within 10% of their Ca2+ load (8.3 ± 1.9% for Fura-2 and 6.7 ± 3.3% for Fura-FF). Antisense-treated neurons recovered to 45% of their Ca2+ load (44.6 ± 5.0% for Fura-2 and 45.7 ± 14.3% for Fura-FF), which is significantly elevated over missense-treated neurons (p < 0.001).
To determine whether [Ca2+]i in antisense-treated neurons eventually returned to basal levels after the glutamate-induced Ca2+ load, neurons were imaged 2 h after glutamate exposure. Missense-treated and antisense-treated neurons were exposed to glutamate (50 μM, 2 min). One hour later, neurons were loaded with Fura-2 for 30 min and then washed three times with recording solution and incubated an additional 15 min for esterase cleavage. Neurons were then placed on the microscope, and 340/380 ratio values were obtained. Two hours after glutamate exposure, ratio values from missense-treated neurons (n = 21) had dropped to 0.24, whereas ratio values from antisense-treated neurons (n = 30) had dropped to 0.31 (Fig. 6). To ensure that basal [Ca2+]i was consistent with what was previously reported, basal [Ca2+]i was obtained from the same batches of neurons and were not statistically different from the values obtained 2 h post-glutamate. However, analysis with the Student's t test revealed that the 2-h post-glutamate values from antisense-treated neurons were still significantly elevated over missense-treated neurons (p < 0.001).
The effects of recovery of CaMK-II suppression on the ability of the neurons to handle a Ca2+ load were also evaluated. Neuronal cultures were treated with αCaM kinase II antisense oligonucleotide for 3 days and then allowed to recover for 3 days, thereby restoring CaMK-II function. Neurons were loaded with Fura-2, and [Ca2+]i was measured every 30 s. A Ca2+ load was produced by brief exposure to glutamate (50 μM, 2 min) (Fig. 7). To compare the Ca2+ decay curves of missense-treated, antisense-treated, and antisense-recovered neurons, the ratio values were normalized to percentage of Ca2+ load as described previously. As reported previously, missense-treated neurons declined to 8.3% of their Ca2+ load, whereas antisense-recovered neurons only recovered to 44.6% of their Ca2+ load during the 30 min of recording. Antisense-recovered neurons recovered to 13.7 ± 2.3% of their Ca2+ load, which is significantly lower than antisense-treated neurons and not significantly different from missense-treated neurons (Fig. 8, P < 0.001, RM ANOVA). By allowing cultures to recover from the antisense oligonucleotide-dependent suppression of CaMK-II, neurons were able to completely buffer [Ca2+]i after glutamate exposure and restore basal Ca2+ levels. The ability of the antisense-recovered neurons to buffer the Ca2+ load indicates that both the development of SREDs and the altered Ca2+ homeostasis were reversible. These results further indicate a relationship between the effects of CaMK-II activity on Ca2+ homeostatic mechanisms and the development of epileptiform activity.
To exclude the possibility that ongoing SRED activity was accounting for observed changes in [Ca2+]i homeostasis, TTX was added to the cultures before [Ca2+]i imaging to inhibit basal electrophysiological activity and epileptiform discharges. Addition of TTX lowered basal [Ca2+]i in both missense-treated and antisense-treated cultures (0.23 and 0.26, respectively). However, with the abolishment of seizure and background activity, basal [Ca2+]i levels from antisense-treated cultures still remained significantly elevated above [Ca2+]i from missense-treated cultures (Fig. 8A; P < 0.001, Student's t test). Likewise, in the presence of TTX, the ability of antisense-treated neurons to recover from a glutamate-induced Ca2+ load is altered compared with missense-treated neurons (Fig. 8B). Antisense-treated neurons recovered to 40.1 ± 7.2% of the Ca2+ load, whereas missense-treated neurons recovered to 11.1 ± 4.6% of the Ca2+ load. Thus, alterations in the [Ca2+]i homeostatic mechanisms associated with decreased CaMK-II activity and protein expression were not solely dependent on the presence of SREDs in this preparation.
Alterations in Ca2+ homeostasis and CaMK-II levels have been observed in various models of epilepsy (Bronstein et al., 1993; Butler et al., 1995; DeLorenzo et al., 1998, 2005; Blair et al., 1999; Churn et al., 2000a). Because epileptogenesis has been associated with increased [Ca2+]i and alterations in Ca2+ homeostatic mechanisms (reviewed in Delorenzo et al., 2005), it was important to determine whether changes in CaMK-II activity were associated with changes in Ca2+ homeostasis and the development of epileptogenesis. The present study provides evidence that suppression of CaMK-II activity is associated with both alterations in Ca2+ homeostasis and the development of SREDs in a reversible manner. Fluorescence imaging of [Ca2+]i, with both Fura-2 and Fura-FF, was used to determine [Ca2+]i in neurons treated with an antisense oligonucleotide to knock down levels of CaMK-II. Addition of antisense oligonucleotide decreased CaMK-II activity and immunocytochemical expression and resulted in the development of SREDs, a characteristic of epilepsy. In the antisense-treated neurons, basal [Ca2+]i was elevated over both control and missense-treated neurons. Reversal of the CaMK-II knockdown following removal of the oligonucleotide restored CaMK-II activity and basal [Ca2+]i and abolished epileptiform activity (SREDs), demonstrating a cause and effect relationship between CaMK-II levels and the development of SREDs and changes in basal [Ca2+]i. [Ca2+]i in antisense-treated neurons were still elevated over [Ca2+]i in missense-treated neurons incubated in TTX to abolish both basal synaptic and seizure activity, demonstrating that CaMK-II changes in [Ca2+]i was effecting the underlying Ca2+ homeostatic mechanisms that regulate basal [Ca2+]i independent of neuronal activity.
In addition to evaluating the effects of CaMK-II knock-down on [Ca2+]i, this study evaluated the effects of altering CaMK-II activity on Ca2+ homeostatic mechanisms that regulate the neuron's ability to handle a Ca2+ load. After a glutamate-induced Ca2+ load, antisense-treated neurons were not able to buffer [Ca2+]i back to basal levels as efficiently as control neurons, similar to findings observed in hippocampal neuronal culture models of both SE and stroke-induced AE (Pal et al., 1999; Sun et al., 2004). Abolishing both basal synaptic and seizure activity with TTX did not affect the altered Ca2+ homeostatic mechanisms observed with decreased CaMK-II activity. To determine the reversibility of the CaMK-II knockdown-induced changes on Ca2+ homeostatic mechanisms and the development of SREDs, antisense-treated neurons were allowed to recover from the antisense oligonucleotide treatment. SREDs were abolished, Ca2+ homeostasis was restored, and recovered neurons were as effective as controls in their ability to effectively buffer [Ca2+]i back to basal levels after a glutamate-induced Ca2+ load. The results from this study employed both the high and lower affinity Ca2+ indicators Fura-2 and Fura-FF. Both missense-treated and antisense-treated neurons loaded with each indicator showed similar responses to the glutamate-induced Ca2+ load. The data suggest a correlation between the suppression of CaMK-II activity with alterations in Ca2+ homeostatic mechanisms and expression of SREDs in hippocampal neurons in culture, indicating that CaMK-II is playing a role in regulating Ca2+ homeostasis during epileptogenesis in this model of epileptiform activity.
A decrease in function of CaMK-II has been shown to occur in a number of models of epilepsy (Bronstein et al., 1993; Butler et al., 1995; Churn et al., 2000a,b). However, this is the first study that shows a direct relationship between a specific suppression of CaMK-II activity and a loss in the ability of neurons to maintain Ca2+ homeostatic mechanisms in association with the development of SREDs. CaMK-II is a ubiquitous enzyme that phosphorylates and regulates the activity of many receptors, including N-methyl-d-aspartate, GABAA, and inositol triphosphate, all of which play a role in epilepsy (Cardy and Taylor, 1998; Huang et al., 2005). CaMK-II activation has been associated with increased excitability, including long-term potentiation of excitatory synapses (reviewed by Bronstein et al., 1993). However, decreased CaMK-II activity has been consistently found in various models of epilepsy. The present study shows that decreasing the expression of CaMK-II results in a hyperexcitable state in neurons in culture. Further studies are needed to determine the role of decreased CaMK-II activity in inducing SREDs and altering Ca2+ homeostasis.
It can be deduced from the results of the present study that the phosphorylation of CaMK-II substrates might be linked to the observed alterations in Ca2+ homeostasis. Further studies need to be conducted to isolate the specific substrate(s) involved in these processes. Interestingly, unlike the SE-induced model of AE (Sombati and Delorenzo, 1995), knocking down levels of CaMK-II does not permanently alter Ca2+ homeostasis or permanently induce SREDs. Our results indicate that a 3-day recovery period is sufficient to fully restore CaMK-II activity, abolish SREDs, and restore the ability of the neurons to regulate Ca2+ homeostasis. Thus, this model of reversible AE may provide unique opportunities to identify the important cellular mechanisms mediating the induction and maintenance of AE and indicate that the effects of CaMK-II in regulating Ca2+ homeostasis plays an important role in the development of AE in hippocampal neurons in culture. Although using cultured hippocampal neurons is a well established model for studying epilepsy, it is important to confirm the results of studies in culture with in vivo models. Further studies are needed to determine the in vivo effects of CaMK-II knockdown to understand the role of CaMK-II in epilepsy.
Calcium plays a pivotal role in normal neuronal function (Llinas et al., 1992; Berridge, 1998). Normal neuronal [Ca2+]i is maintained around 100 nM (Mody and MacDonald, 1995). This concentration is less than one-ten thousandth of the free extracellular Ca2+ concentration ([Ca2+]e) (Putney, 1999). In light of the important signaling function of Ca2+ and the excitotoxic implications of excess [Ca2+]i, neurons have an intricate system to regulate [Ca2+]i. Evidence indicates that Ca2+ homeostasis is markedly altered acutely during the injury and epileptogenic phases of AE (Raza et al., 2004) and even in the chronic spontaneous recurrent seizures (epilepsy) phases of AE (Pal et al., 2000, 2001; Raza et al., 2001, 2004). These observations provide direct evidence that the alterations of normal Ca2+ homeostasis plays a role in epileptogenesis and the maintenance of the epileptic phenotype (Delorenzo et al., 2005).
Complex regulatory processes are mediated by Ca2+ homeostatic mechanisms in neurons (Carafoli et al., 1997). Increased or prolonged entry of extracellular Ca2+ could contribute to the altered Ca2+ homeostatic mechanisms in epilepsy (Carafoli et al., 1997). [Ca2+]i homeostasis is regulated by a number of cellular mechanisms that include calcium-induced calcium release from intracellular stores, sequestration into the endoplasmic reticulum by the sarco/endoplasmic reticulum calcium ATPase (SERCA) (Carafoli et al., 1997), mitochondrial uptake of Ca2+ (Kunz et al., 1999), and efflux via the plasma membrane Na+/Ca2+ exchanger (Ryan, 1999). In addition, a number of [Ca2+]i buffering systems that include the binding proteins calbindin, calretinin, and parvalbumin (Nagerl et al., 2000) are important regulators of [Ca2+]i. In the CaMK-II knockdown neurons, Ca2+ homeostasis was disrupted as indicated by elevated basal [Ca2+]i and the impaired ability to buffer a Ca2+ load. These results are similar to previous observations of altered [Ca2+]i in other models of epilepsy, including the low Mg2+-induced SE model of AE (Pal et al., 1999) and the stroke model of AE (Sun et al., 2004). Studies have implicated decreased activity of SERCA leading to increases in [Ca2+]i in whole animal and in vitro models of epilepsy (Parsons et al., 2000; Pal et al., 2001). Enhanced calcium-induced calcium release via the inositol triphosphate receptor has also been shown to be a contributor to increased [Ca2+]i and altered Ca2+ homeostatic mechanisms in epilepsy (Pal et al., 2001). Further studies are needed to evaluate what cellular pathway(s) is involved in the loss of Ca2+ homeostasis in the CaMK-II antisense knockdown model of AE.
CaMK-II activity plays a role in maintaining Ca2+ homeostasis in both skeletal and cardiac muscle, with a positive correlation between CaMK-II activity and SERCA activity. In these tissues, CaMK-II has been shown to regulate SERCA by two different mechanisms. One mechanism involves CaMK-II phosphorylation of the SERCA accessory protein, phospholamban, which enhances the uptake of Ca2+ by SERCA (Colyer, 1998). Although phospholamban is not present in neurons (Plessers et al., 1991), there is evidence that a protein analogous to phospholamban exists in neurons and may be required for neuronal SERCA activity (Dou and Joseph, 1996). Furthermore, SERCA activity in skeletal muscle is increased by direct phosphorylation by CaMK-II (Xu et al., 1993). This relationship has yet to be studied in neurons and could provide the link between decreased CaMK-II activity and altered Ca2+ homeostasis in our model of epilepsy.
A better understanding of the molecular mechanisms that underlie the relationship between altered Ca2+ dynamics and the induction and maintenance of AE is important for understanding the process of epileptogenesis (Delorenzo et al., 2005). Although further studies are needed, it is apparent that elucidating the molecular basis contributing to altered Ca2+ regulatory mechanisms in the CaMK-II knockdown model of AE may provide an insight into the long-term plasticity changes associated with epilepsy and offer specific molecular targets for preventing and possibly reversing the pathophysiological condition of AE.
This work was supported by a National Institutes of Health, National Institute of Neurological Disorders and Stroke Grants R01-NS23350 (to R.J.D.) and P50-NS25630 (to R.J.D.) and the Milton L. Markel Alzheimer's Disease Research and the Sophie and Nathan Gumenick Neuroscience Research Funds. D.S.C. and S.N.H. are supported by the National Institutes of Health Training Grant T32-NS007288-20.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
ABBREVIATIONS: SREDs, spontaneous recurrent epileptiform discharges; AE, acquired epilepsy; SE, status epilepticus; CaMK-II, calcium/calmodulin-dependent protein kinase II; DIV, days in vitro; PIPES, piperazine-N,N′-bis(2-ethanesulfonic acid); pBRS, physiological bath recording solution; SERCA, sarco(endo)plasmic reticular calcium ATPase; TTX, tetrodotoxin; PBS, phosphate-buffered saline; RM, repeated measures; ANOVA, analysis of variance.
- Received July 5, 2006.
- Accepted September 11, 2006.
- The American Society for Pharmacology and Experimental Therapeutics