SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix

Transient, non-catastrophic brain ischaemia can induce either a protected state against subsequent episodes of ischaemia (ischaemic preconditioning) or delayed, selective neuronal death. Altered glutamatergic signalling and altered Ca2+ homeostasis have been implicated in both processes. Here we use simultaneous patch-clamp recording and Ca2+ imaging to monitor early changes in glutamate release and cytoplasmic [Ca2+] ([Ca2+]c) in an in vitro slice model of hippocampal ischaemia. In slices loaded with the Ca2+-sensitive dye Fura-2, ischaemia leads to an early increase in [Ca2+]c that precedes the severe ischaemic depolarization (ID) associated with pan necrosis. The early increase in [Ca2+]c is mediated by influx through the plasma membrane and release from internal stores, and parallels an early increase in vesicular glutamate release that manifests as a fourfold increase in the frequency of miniature excitatory postsynaptic currents (mEPSCs). However, the increase in mEPSC frequency is not prevented by blocking the increase in [Ca2+]c, and the early rise in [Ca2+]c is not affected by blocking ionotropic and metabotropic glutamate receptors. Thus, the increase in [Ca2+]c and the increase in glutamate release are independent of each other. Stabilizing actin filaments with jaspamide or phalloidin prevented vesicle release induced by ischaemia. Our results identify several early cellular cascades triggered by ischaemia: Ca2+ influx, Ca2+ release from intracellular stores, actin filament depolymerization, and vesicular release of glutamate that depends on actin dynamics but not [Ca2+]c. All of these processes precede the catastrophic ID by several minutes, and thus represent potential target mechanisms to influence the outcome of an ischaemic episode.

Abbreviations 
d-AP5

d-(–)-2-amino-5-phosphonopentanoic acid

ID

ischaemic depolarization

(S)-MCPG

(S)-α-methyl-4-carboxyphenylglycine

mEPSC

miniature excitatory postsynaptic current

NBQX

2,3-dioxo-6-nitro-1,2,3,4-tetrahydrobenzo[f]quinoxaline-7-sulfonamide

TTX

tetrodotoxin

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix

Ischaemia-induced brain damage is a leading cause of death and disability. In vivo studies have identified two main phases of response to severe brain ischaemia (Astrup et al. 1977; Hansen & Nedergaard, 1988; Rossi et al. 2007). The first phase, lasting 1–2 min, is characterized by an isoelectric EEG and a gradual elevation of extracellular K+, with little change in the concentration of other principal ions or electrical potential. The second phase is characterized by a rapid and large ischaemic depolarization (ID), loss of principal ionic gradients, and build up of glutamate and other neurotransmitters in the extracellular space. If the latter phase persists for more than a couple of minutes, pan necrosis ensues. In contrast, shorter transient episodes of ischaemia can lead either to ischaemic preconditioning, whereby brain tissue becomes more resistant to subsequent episodes of ischaemia (Obrenovitch, 2008), or to selective cell death in ischaemia-sensitive neurons (Pulsinelli, 1985). While much is known about how the ID and release of transmitters damages tissue during the second phase of ischaemia (Lipton, 1999; Rossi et al. 2007), less is known about the earlier events of ischaemia and their relationship to preconditioning or transition to selective neuronal death.

In vitro brain slice studies have determined that one of the earliest manifestations of brain ischaemia/anoxia is an increase in the frequency of miniature excitatory postsynaptic currents (mEPSCs) (Fleidervish et al. 2001; Katchman & Hershkowitz, 1993; Allen et al. 2004). This early increase in glutamatergic excitation could contribute either to ischaemic preconditioning (Jiang et al. 2003; Miao et al. 2005; Lin et al. 2008), or to selective neuronal death in ischaemia-sensitive neurons (Gill et al. 1987).

An early in vivo study using Ca2+-sensitive electrodes detected an early, small rise in [Ca2+]c in hippocampal CA1 neurons that precedes the ID (Silver & Erecinska, 1990), raising the possibility that increased mEPSC frequency is due to elevated presynaptic [Ca2+]c. However, in a study of hippocampal slices, the anoxia-induced increase in mEPSC frequency was independent of both action potentials and external Ca2+ (Katchman & Hershkowitz, 1993). In the same study, dantrolene, a ryanodine receptor antagonist, prevented the increase in mEPSC frequency, leading the authors to suggest that the increased mEPSC frequency is triggered by Ca2+ release from intracellular stores in presynaptic terminals. However, this conclusion was complicated somewhat by an effect of dantrolene on basal mEPSC frequency, and by lack of information on the Ca2+ dynamics during anoxia in these experiments. Thus, it remains uncertain whether the ischaemia-induced increase in mEPSC frequency is triggered by Ca2+.

We used patch-clamp recording with simultaneous Ca2+ imaging in hippocampal slices during simulated ischaemia. Ischaemia caused an early increase in [Ca2+]c, mediated by Ca2+ influx and release from intracellular stores, that paralleled the increase in mEPSC frequency. Surprisingly, preventing the rise in [Ca2+]c by removing external Ca2+ and soaking slices in BAPTA-AM, did not affect the increase in mEPSC frequency. The increase in mEPSCs was, however, prevented by pre-treating slices with the actin filament stabilizers jaspamide and phalloidin oleate suggesting that ischaemia-induced vesicle release is mediated by actin filament depolymerization in presynaptic terminals.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix

Ethical information

The authors have read The Journal of Physiology article on reporting of ethical standards (Drummond, 2009), and all of our procedures comply with their policies and regulations, as well as to the regulations detailed in the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and were approved by the Animal Care and Use Committee of the Oregon Health and Science University.

Preparation of brain slices

Seventy Sprague–Dawley rats contributed to the present study. The animals were housed with ad libitum access to food and water in a room air-conditioned at 22–23°C with a standard 12 h light–dark cycle. The hippocampus was obtained from Sprague–Dawley albino rats (Rossi & Slater, 1993; Rossi & Hamann, 1998; Rossi et al. 2000). Rats (18–21 days old) were anaesthetized with isoflurane (administered by inhalation), and killed by decapitation. The whole brain was rapidly isolated and immersed in ice cold (0–2°C) artificial cerebrospinal fluid (ACSF) containing (in mm): 124 NaCl, 26 NaHCO3, 1 NaH2PO4, 2.5 KCl, 2.5 CaCl2, 2 MgCl2, 10 d-glucose, and bubbled with 95% O2–5% CO2 (pH 7.4). The hippocampus was dissected out of the brain and mounted in a slicing chamber filled with ice cold (0–2°C) ACSF. Slices (225 μm) were made with a vibrating tissue slicer (Vibratome). Slices were incubated in warmed ACSF (33 ± 1°C) for 1 h after dissection and then held at 22–23°C until used. Kynurenic acid (1 mm) was included in the dissection, incubation and holding solution (to block glutamate receptors to reduce potential excitotoxic damage) but was omitted from the experimental solutions.

Patch-clamp recording of glutamatergic mEPSCs from hippocampal pyramidal cells

Slices were placed in a submersion chamber on an upright microscope, and viewed with an Olympus 60× (0.9 numerical aperture) water immersion objective with differential interference contrast and infrared optics. Slices were perfused with heated (32–34°C) ACSF at a rate of ∼4 ml min−1. Drugs were dissolved in ACSF and applied by bath perfusion. Whole-cell recordings were made from the somata of visually identified CA1 pyramidal cells. Patch pipettes were constructed from thick-walled borosilicate glass capillaries and filled with an internal solution containing (mm): 130 Caesium gluconate, 4 NaCl, 0.5 CaCl2, 10 Hepes, 5 EGTA, 4 MgATP, 0.5 Na2GTP, 5 QX-314 (to suppress voltage-gated sodium currents), pH adjusted to 7.2 with CsOH. Electrode resistance was 1.5 to 2.5 MΩ. Cells were rejected if access resistance was greater than 5 MΩ. Cells were also rejected if the access resistance, monitored with −5 mV voltage steps, changed by more than 20% during the course of an experiment. To isolate glutamatergic mEPSCs all recordings were done in ACSF supplemented with tetrodotoxin (0.5–1 μm) and the GABAA antagonist GABAzine (10 μm).

Simulating ischaemia in brain slices

We simulated severe brain ischaemia by exposing slices to solution in which glucose and oxygen were replaced with sucrose and nitrogen, and supplemented with iodoacetic acid (2 mm) to block glycolysis. Previously we also supplemented our ischaemia simulation solution with cyanide (1 mm) to block oxidative phosphorylation (Rossi et al. 2000; Hamann et al. 2005). Because cyanide may have chemical interactions with either glutamate receptors or glutamate transporters independent of cellular responses to energy deprivation, in this project we did not include cyanide. The only significant difference that we observed between these two methods of simulating ischaemia was a slightly more rapidly developing response when cyanide was included (data not shown).

Loading slices with Fura-2-AM and BAPTA-AM

Stock solutions of the acetoxymethyl (AM) ester of Fura-2 (10 mm) or BAPTA (20 mm) were made in DMSO with 20% pluronic acid, and sonicated for ∼1 min on the day of experiments. Stock solutions were added to our standard incubation solution at final concentrations of 100 μm for Fura-2-AM and 200 μm BAPTA-AM, followed by brief sonication. Slices were incubated in final solution for 1 h, and experiments were conducted within 45 min of removal from the incubation solution.

Data acquisition, analysis and statistics

CA1 pyramidal cells were voltage-clamped at −60 mV with an Axoclamp 700b patch clamp amplifier (Molecular Devices, Sunnyvale, CA, USA), and mEPSCs were filtered at 5 kHz and acquired at 20 kHz with pCLAMP software (Molecular Devices). Synaptosoft mini analysis software (Synaptosoft Inc., Decatur, GA, USA) was used to analyse mEPSC frequency and amplitude. Fura-2 was excited at 340 and 380 nm light with a Lambda DG4 ultra high speed wavelength switcher (Sutter Instrument Co., Novato, CA, USA). Whole field (1343 × 780 pixels after 600× magnification) Fura-2 fluorescence emission was acquired with an ORCA-ER digital camera (Hamamatsu) and analysed using Slidebook software (Olympus). All data are presented as means ±s.e.m. To determine if a given treatment affected mEPSC frequency over time in ischaemia, a repeated measures two-way ANOVA with Bonferroni's post hoc test was used, and statistical significance was defined as P < 0.05. Student's t test for unpaired data was used to determine the effect of treatments on baseline parameters, such as EPSC amplitude and frequency.

Quantifying changes in Fura-2 fluorescence emission

All values are expressed as a percentage change in the ratio of Fura-2 fluorescence emission when excited at 340 nm and 380 nm with the equation (FtF0)/F0, where Ft is the emission ratio at time t and F0 is the emission ratio at time 0. For all experimental conditions, a separate, parallel series of experiments was conducted with slices that were not loaded with Fura-2, but were otherwise treated the same, and the autofluorescence emission at each excitation wavelength was averaged for three to six slices (Fig. 2). The resultant mean autofluorescence signal for each excitation wavelength was subtracted from the experimental emission signal for each individual Fura-2 loaded slice (Fig. 2). The autofluorescence subtraction procedure did not alter the experimental outcomes or conclusions because autofluorescence did not change dramatically during the early phases of ischaemia in which all experiments were conducted (before the ischaemic depolarization). Furthermore, although the various 0 Ca2+ experimental treatments did affect baseline autofluorescence magnitude (no other experimental treatments affected baseline autofluorescence), changes in autofluorescence induced by ischaemia were minimal for all experimental treatments (Fig. 2). For 0 Ca2+ experiments, because switching to 0 Ca2+ Ringer solution resulted in a gradual but continuous decline in [Ca2+]c, control experiments were conducted to obtain an average waveform of the Fura-2 emission ratio upon switching to 0 Ca2+ Ringer solution. The resultant mean Fura-2 ratio waveform was subtracted from each slice exposed to ischaemia in 0 Ca2+ to isolate the ischaemia-induced response.

image

Figure 2. Autofluorescence subtraction Aa–Ca, traces show plots of Fura-2 emission ratios calculated before (open circles) and after (filled circles) subtraction of mean autofluorescence values at 340 and 380 nm excitation, as derived from parallel experimental slices that were not soaked in Fura-2-AM, shown in traces Ab–Cb (scale bar for autofluorescence traces is in arbitrary units of fluorescence emission intensity (f.e.i.), and the 0 f.e.i. level is marked by an arrowhead in Ab which applies to AbCb). Autofluorescence data and resultant subtraction plots are for control data (A, used in Fig. 3C), 0 Ca2+ data (B, used in Fig. 5A), and 0 Ca2+ after BAPTA-AM pre-soak (C, used in Fig. 7A). Note minimal changes in autofluorescence during the time period of ischaemia that we analyse (Ab–Cb), with resultant minimal effect of autofluorescence subtraction on calculated ratio plots (Aa–Ca). Also note the small error bars for autofluorescence plots (Ab–Cb) indicating consistency across slices. This procedure was carried out for all Ca2+ imaging experiments, with n= 3–6 autofluorescence slices for each experimental condition.

Download figure to PowerPoint

Reagents

Fura-2-AM (Invitrogen, Carlsbad, CA, USA), BAPTA-AM (Sigma, St Louis, MO, USA), AP5, TTX, NBQX, (S)-MCPG (Ascent Scientific Ltd, Avonmouth, Bristol, UK), jaspamide (Alexis Biochem, Plymouth Meeting, PA, USA), phalloidin (Calbiochem).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix

Ischaemia increases [Ca2+]c and mEPSC frequency in parallel

To determine the influence of ischaemia on [Ca2+]c and mEPSC frequency in the hippocampus, we pre-incubated hippocampal slices with the membrane permeant Ca2+ indicator Fura-2-AM, and voltage-clamped CA1 pyramidal cells in the Fura-2 loaded slices (Fig. 1A and B). We used Fura-2-AM because it has a high affinity for Ca2+ (Kd= 145 nm), enabling us to detect small tonic changes in [Ca2+]c with better resolution than lower affinity indicators such as rhod-2-AM (Mitani et al. 1993) (Kd= 570 nm), and because of potential movement or tissue swelling which can occur during ischaemia, it was important to use a ratiometric dye. To ensure that any changes in Fura-2 fluorescence that we observe are due to changes in [Ca2+]c, and not changes in NADH autofluorescence (Brooke et al. 1996), for all experimental conditions we have quantified and subtracted autofluorescence signals mediated by NADH and other endogenous fluorophores from the Fura-2 signal (Fig. 2). Because the slices are bulk-loaded and imaged non-confocally, it is not possible to monitor individual cellular compartments. Accordingly, we quantified [Ca2+]c as the total fluorescence emission intensity across the entire field of view, which for all experiments had identical dimensions and was centred over the CA1 pyramidal cell layer (as in Fig. 1A and B).

image

Figure 1. Simulated ischaemia causes an early rise in [Ca2+]c and an increase in mEPSCs prior to the ischaemic depolarization A, DIC image of patch-clamped CA1 pyramidal cell in a hippocampal slice that has been bulk loaded with the Ca2+ sensitive dye Fura-2-AM. B, fluorescence emission image of the Fura-2-AM loaded slice shown in A, excited at 340 nm. Scale bar is 15 μm, and applies to images in A and B. C, whole field fluorescence emission ratio (grey, top trace, expressed as percentage change) and the corresponding membrane current (black, bottom trace) of a voltage-clamped (Vh=−60 mV) CA1 pyramidal cell, in the field of fluorescence measurement shown in A and B. In this and all subsequent experiments, the value used for deriving, plotting and analysing ratios is the total fluorescence emission from the entire field of view shown in B. For all experiments the dimensions of the field of view are identical, and are centred on the pyramidal cell layer as in B. Note, + and ++ indicate regions of membrane current used for expanded time and amplitude display of mEPSCs in D. D, mEPSCs in expanded time scale plot of membrane current from the regions of the macroscopic current indicated by the + corresponding to control, and during the early stage of simulated ischaemia ++, prior to the ID current. Individual mEPSCs are indicated by asterisks. Note, the 4 sweeps in each set are temporally contiguous. In all subsequent experiments, the analysis of mEPSCs and [Ca2+]c are restricted to the time period preceding the ID shown in C.

Download figure to PowerPoint

Figure 1C shows the effects of solutions designed to mimic severe brain ischaemia (in vitro ischaemia, see Methods for details) on [Ca2+]c in the CA1 region and the simultaneously recorded macroscopic current of a pyramidal cell located within the field of view being imaged. Ischaemia caused an early, gradual increase in [Ca2+]c, (Fig. 1C, top trace, time referred to marked by ++) followed after several minutes by a rapid, large increase in [Ca2+]c which corresponds to the large inward current recorded simultaneously in a CA1 pyramidal cell located within the field being imaged (Fig. 1C, bottom trace). Our previous work has shown that the large inward current is generated by glutamate receptor-gated channels (Rossi et al. 2000), which in vivo studies indicate underlies the large increase in [Ca2+]c (Silver & Erecinska, 1990). Although there was no major change in pyramidal cell macroscopic membrane current during the early rise in [Ca2+]c (marked by ++ in Fig. 1C), there was a gradual increase in the frequency of mEPSCs (Fig. 1D) that closely paralleled the early rise in [Ca2+]c (Fig. 3AC; mean mEPSC frequency after 5 min in ischaemia was 1.32 ± 0.81 Hz, n= 8, significantly greater, P < 0.05, than baseline frequency of 0.35 ± 0.19, and Fura-2 emission ratio at 5 min in ischaemia was 27.9 ± 8.0% greater than baseline ratio, P < 0.05, n= 8). While mEPSC frequency increased approximately fourfold, the mean amplitude was not affected (Fig. 3D and E).

image

Figure 3. Ischaemia-induced increase in mEPSCs parallels ischaemia-induced rise in [Ca2+]c A, cumulative fraction plot of mEPSC inter-event interval for a single cell, under control conditions (black filled circles) and during the early stages of ischaemia (grey filled circles). B, plot of mEPSC frequency (in 10 s bins) aligned with mean percentage change in Fura-2 fluorescence ratio for a single experiment (black line). C, plot of mean mEPSC frequency aligned with mean percentage change in Fura-2 fluorescence ratio (n= 8). The 0 Hz level is marked by the arrowhead. D, cumulative fraction plot of mEPSC amplitude for a single cell, under control conditions (black filled circles) and during the early stages of ischaemia (grey filled circles). Inset shows mean mEPSC waveform aligned by rise time in control (black) and in ischaemia (grey). E, plot of mEPSC amplitude under control conditions versus mEPSC amplitude during ischaemia. F, plot of mean ischaemia-induced increase in mEPSC frequency in cells from slices that were: loaded with Fura-2-AM and exposed to UV light as in B and C (black filled circles), loaded with Fura-2-AM without UV exposure (grey filled circles), or neither loaded with Fura-2-AM nor exposed to UV light (open circles). The mean mEPSC frequency in Fura-2 loaded slices or Fura-2 loaded and UV exposed slices was not significantly different from the mEPSC frequency in unloaded, unexposed slices at any time point during control or ischaemia. The 0 Hz level is marked by the arrowhead.

Download figure to PowerPoint

The increase in mEPSC frequency was not affected by either Fura-2 loading or by imaging with UV light (Fig. 3F), demonstrating that our imaging procedure did not have spurious effects on spontaneous or ischaemia-evoked glutamate release. Thus, simulated ischaemia causes an early, parallel increase in [Ca2+]c and mEPSC frequency that precedes by several minutes the ID and massive glutamate release that causes pan necrosis. In this and all subsequent data sets, we restrict our analysis of mEPSCs and [Ca2+]c to the time period preceding the large ID current and associated large increase in [Ca2+]c shown in Fig. 1C.

Before determining the relationship between the early rise in [Ca2+]c and mEPSC frequency, we wanted to ensure that the Fura-2 signals we record would enable us to detect changes in [Ca2+]c in presynaptic terminals (in addition to other cellular compartments), should they occur. To do this we electrically stimulated the Schaeffer collateral axon track, recorded the corresponding changes in Fura-2 fluorescence ratio over the entire field of view (as shown for ischaemia in Fig. 1B), and pharmacologically identified the subcellular makeup of the resultant global Ca2+ signal (Fig. 4A and C). Tetanic stimulation of the Schaeffer collaterals (400 μA, 50 Hz, 4 s) evoked summating EPSCs (Fig. 4B) and an associated transient increase in [Ca2+]c, as evidenced by an increase in the Fura-2 fluorescence ratio (Fig. 4A). Bath application of ionotropic glutamate receptor antagonists (d-AP5, 50 μm; NBQX, 25 μm) abolished the summating EPSCs (Fig. 4B) but only partially reduced the amplitude of the Ca2+ transient (Fig. 4A and C; Fura-2 emission ratio was 18.6 ± 7.1% of control, n= 7); similarly, a residual Ca2+ signal persisted in the presence of both these ionotropic glutamate receptor antagonists and the broad spectrum metabotropic glutamate receptor (mGluR) antagonist (S)-MCPG (500 μm) (Fig. 4C; Fura-2 emission ratio was 29.3 ± 7.8% of control, n= 6). The residual Ca2+ signal, when glutamate receptors were blocked, was abolished by the sodium channel antagonist TTX (0.5 μm) (Fig. 4A and C; Fura-2 emission ratio was 0.6 ± 0.3 and −1.9 ± 2.5% of control, n= 7 and 6, respectively, for without or with MCPG, both significantly reduced compared to glutamate antagonists alone, P < 0.05) suggesting that it reflects action potential-evoked elevations of [Ca2+]c in presynaptic terminals and axons. However, before concluding that the tetanization-evoked, glutamate receptor-resistant, TTX-sensitive 20–30% of the Fura-2 signal (Fig. 4C) reflects changes of [Ca2+]c in presynaptic terminals and axons, we tested whether the cocktail of glutamate receptor antagonists used for the bulk-loaded slice experiments actually abolished postsynaptic neuronal and astrocytic responses to tetanic stimulation (Fig. 4D and E). We loaded individual CA1 pyramidal neurons and astrocytes with Fura-2 in the recording electrode (Fig. 4Da and Ea), and used current clamp recording with physiological internal solutions to allow the cells to behave as they would in bulk loaded slices. Under these conditions, the tetanic stimulation protocol evoked summating EPSPs which triggered overshooting action potentials in pyramidal cells (Fig. 4Db, bottom, n= 3), but no detectable electrical response in astrocytes (Fig. 4Eb, bottom, n= 3). The tetanic stimulation also elevated [Ca2+]c in the pyramidal cell soma, axon and dendrites (Fig. 4Db, top), but had no affect on astrocyte [Ca2+]c (Fig. 4Eb, top). Both the electrical and Ca2+ response of the pyramidal cell were abolished by d-AP5 + NBQX (Fig. 4Dc, n= 3). Since tetanization does not elicit detectable changes in neuronal or astrocyte [Ca2+]c in the presence of d-AP5 and NBQX, we conclude that the d-AP5, NBQX and MCPG resistant, but TTX sensitive, tetanization-induced Fura-2 signal in bulk loaded slices reflects elevations of [Ca2+]c in presynaptic terminals and axons. Thus, for our ischaemia experiments, the Fura-2 signal in our bulk loaded slice experiments can detect changes in [Ca2+]c if they occur in either pre- or postsynaptic compartments.

image

Figure 4. Fura-2 signal in bulk loaded slices reflects [Ca2+]c in pre- and postsynaptic compartments A, ratio of the whole field fluorescence emission at 340/380 nm excitation during trains of stimulation under various pharmacological conditions. The continuous bar indicates application of AP5 (50 μm), NBQX (25 μm) and MCPG (500 μm), and the dashed bar indicates application of TTX (0.5 μm). Arrows indicate time of tetanic stimulation. B, summating EPSC in a voltage-clamped (Vh=−60 mV) CA1 pyramidal cell, evoked by tetanizing the Schaeffer collateral pathway (black), is abolished by AP5 and NBQX (dark grey) or AP5, NBQX and TTX (light grey). Inset shows an expanded time frame of the rising phase of the response, allowing visualization of individual EPSCs and complete block by AP5 and NBQX. Arrowheads indicate individual stimuli. C, bar chart depicting the mean of the peak percentage change of fluorescence emission ratio under various pharmacological conditions (expressed as a percentage of the peak response under control conditions). Conditions are d-AP5 (50 μm) and NBQX (25 μm) either alone (left, white) or combined with (S)-MCPG (500 μm, left, grey), or d-AP5, NBQX and TTX (0.5 μm) either alone (right, white), or combined with (S)-MCPG (right, grey). Da shows a pyramidal cell loaded with Fura-2 via a patch pipette filled with physiological solution. Db, lower trace shows the voltage response of the cell in Da to tetanic stimulation and the top traces show the corresponding changes in Fura-2 fluorescence ratio for the ROIs in Da. Note, ROI 1 is placed on the axon, ROI 2 is placed on a basilar dendrite, ROI 3 is placed on the soma, ROI 4 is placed on the apical dendrite, and ROI 5 is placed on the background (note lack of detectable signal from background). Dc shows that both the voltage response and changes in Fura-2 ratio evoked by tetanization are abolished by d-AP5 (50 μm) + NBQX (25 μm). The traces are for the same cell and ROIs as in Db. Ea shows an astrocyte loaded with Fura-2 via a patch pipette filled with physiological solution. Eb lower trace shows the lack of voltage response of the cell in Ea to tetanic stimulation and the top traces show the corresponding lack of changes in Fura-2 fluorescence ratio for the ROIs in Ea. Note, ROI 1 and 2 are placed on proximal processes and ROI 3 is placed on astrocyte soma. Data shown for pyramidal cell and astrocyte are representative of n= 3 experiments for both cell types.

Download figure to PowerPoint

Removing extracellular Ca2+ reduces the ischaemia-induced increase in [Ca2+]c but not the increase in mEPSCs

A previous study of hippocampal slices showed that anoxia-induced increases in mEPSC frequency persist in nominally Ca2+-free bathing solution (Katchman & Hershkowitz, 1993). However, even trace amounts of extracellular Ca2+ in nominally Ca2+-free solutions might represent a residual source of Ca2+ influx; indeed, we have observed that evoked synaptic responses are reduced but not abolished in nominally Ca2+-free solutions, unless a chelator is added to remove trace Ca2+ (as in Fig. 5C). In extracellular solutions in which we replaced Ca2+ with Mg2+ in the bath and added EGTA (2 mm) to chelate trace residual Ca2+, simulated ischaemia still caused an early rise in [Ca2+]c, but with a significantly reduced magnitude when compared to the rise in Ca2+ seen in control solutions (Fig. 5A; peak increase in Fura-2 emission ratio was 10.7 ± 2.2%, n= 11, with ischaemia causing a significant rise for both conditions, P < 0.001, but with a significant interaction, P < 0.001, between the two conditions, with the points marked with an asterisk being significantly different, P < 0.01). Despite the reduced magnitude of the increase in [Ca2+]c, the increase in mEPSC frequency was not affected (Fig. 5B; n= 8, P > 0.05 for treatment and interaction with ischaemia, compared to ischaemia in normal Ca2+). These data indicate that some of the ischaemia-induced rise in [Ca2+]c is via Ca2+ influx across the plasma membrane; however, the increase in mEPSC frequency is not dependent on this Ca2+ influx.

image

Figure 5. Removing extracellular Ca2+ reduces the ischaemia-induced rise in [Ca2+]c without affecting the increase in mEPSCs A, plot of mean percentage change in Fura-2 emission ratio under normal ischaemia conditions (filled circles) and when extracellular Ca2+ is replaced with equimolar Mg2+ and 2 mm EGTA to chelate trace residual Ca2+ (open circles). Asterisks indicate time points where the percentage increase is significantly different between conditions (P < 0.05, as assessed by Bonferroni's post hoc test on the two conditions which showed a significant interaction with ischaemia, P < 0.001, n= 8 for control and 11 for 0 Ca2+). B, plot of mean mEPSC frequency under normal ischaemia conditions (filled circles) and with extracellular Ca2+ removed (open circles) as in A. The mEPSC frequencies were not significantly different at any time point, and there was no significant interaction between condition and ischaemia (P > 0.05, n= 8 for control and 9 for 0 Ca2+). The 0 Hz level is marked by the arrowhead. C, plot showing mean eEPSC amplitude under control conditions and in 0 Ca2+ saline with 2 mm EGTA added to chelate trace residual Ca2+ (n= 3). Inset shows a representative example of the complete block of the eEPSC by removal of Ca2+ with added EGTA.

Download figure to PowerPoint

Preventing the rise of [Ca2+]c does not prevent the ischaemia-induced increase in mEPSCs

The data in Fig. 5 indicate that a portion of the rise in [Ca2+]c is due to release from intracellular stores, and it has been suggested that Ca2+ release from intracellular stores in presynaptic terminals is the trigger for vesicle release during anoxia (Katchman & Hershkowitz, 1993). To test this idea, we loaded slices with the membrane permeant Ca2+ chelator BAPTA-AM to buffer [Ca2+]c released from intracellular sources (Abdel-Hamid & Tymianski, 1997). To confirm that this would effectively buffer [Ca2+]c, we examined the effects of BAPTA loading on spontaneous and stimulus-evoked EPSCs. Pre-incubation of slices with BAPTA-AM (see Methods for details) abolished stimulus evoked EPSCs at all but the highest stimulus intensities, and even under those conditions the EPSCs were profoundly reduced in amplitude (Fig. 6A; mean eEPSC amplitude at 1 mA stimulation was 123.8 ± 71.6 pA, n= 4, for BAPTA-AM soaked slices, and 494.7 ± 83.8 pA, n= 12, for control slices, P < 0.05). Paired pulse facilitation (PPF) of these residual EPSCs was abolished in BAPTA loaded slices (Fig. 6B, paired pulse ratio was 0.91 ± 0.13, n= 4, for BAPTA-AM soaked slices and 1.31 ± 0.04, n= 7, for control slices, P < 0.05), as expected given that PPF is [Ca2+]c dependent (Regehr et al. 1994; Xu-Friedman & Regehr, 2004). BAPTA loading also significantly reduced the basal mEPSC frequency (Fig. 6C; mEPSC frequency was 0.12 ± 0.02 Hz, n= 8, for BAPTA-AM soaked slices and 0.37 ± 0.12 Hz, n= 8, for control slices, P < 0.05), without affecting mEPSC amplitude (not shown, mEPSC amplitude was 32.9 ± 2.3 pA, n= 8, for BAPTA-AM slices and 32.6 ± 1.3 pA, n= 8, for control slices). Taken together, these data indicate that soaking slices in BAPTA-AM effectively loads presynaptic terminals with BAPTA, reducing basal [Ca2+]c, and strongly buffering Ca2+ transients.

image

Figure 6. BAPTA-AM loading reduces basal [Ca2+] and strongly buffers Ca2+ transients in presynaptic terminals A, plot of eEPSC stimulus intensity response curves for control slices (filled circles) and slices pre-soaked in BAPTA-AM (open circles), both in normal extracellular [Ca2+]. For both conditions, eEPSC amplitudes did not increase for stimulus strengths greater than 1 mA (up to 10 mA). Asterisks indicate stimulus intensities for which the amplitude of the eEPSCs are significantly different (P < 0.05, t test, n= 4). Inset shows representative mean EPSC evoked by 1 mA stimuli in a control slice (black) and a BAPTA loaded slice (dotted). B, plot of paired pulse ratio evoked with maximal stimulus strength (1 mA) for control slices (filled circles) and slices pre-soaked in BAPTA-AM (open circles), both in normal extracellular [Ca2+]. Asterisks indicate significant differences between conditions (P < 0.05, t test, n= 4). C, plot of mean mEPSC frequency in control slices (filled bar) and slices pre-soaked in BAPTA-AM (open bar), both in normal extracellular [Ca2+]. Asterisk indicates that mEPSC frequency was significantly reduced (P < 0.05, t test, n= 8) in BAPTA-AM soaked slices.

Download figure to PowerPoint

We next repeated our ischaemia experiments in the absence of external Ca2+, in slices that were pre-incubated in BAPTA-AM. Under these conditions, the ischaemia-induced increase in [Ca2+]c was abolished (Fig. 7A; with the effects of ischaemia on Fura-2 emission significantly interacting with condition, P < 0.001, and only the last data point exhibiting a significant change compared to control for BAPTA-AM treated slices). However, despite preventing the rise in [Ca2+]c, BAPTA-AM did not prevent the ischaemia-induced increase in mEPSC frequency (Fig. 7B and C; although there was a significant difference in mEPSC frequency between conditions, P < 0.05, there was a significant affect of ischaemia for both conditions, P < 0.001, with no significant interaction, P > 0.8). The lack of effect on the ischaemia-induced increase in mEPSC frequency is highlighted by plotting mEPSC frequencies normalized to the control period for each condition (Fig. 7C).

image

Figure 7. Soaking slices in BAPTA-AM and removing extracellular Ca2+ abolishes ischaemia-induced increases in [Ca2+]c, but does not prevent ischaemia-induced increases in mEPSC frequency A, plot of mean percentage change in Fura-2 emission ratio under normal ischaemia conditions (filled circles) and when BAPTA-AM presoaked slices are exposed to ischaemia with extracellular Ca2+ removed as in Fig. 5 (open circles). Asterisks indicate time points where the percentage increase is significantly different between conditions (P < 0.05, as assessed by Bonferroni's post hoc test on the two conditions which showed a significant interaction with ischaemia, P < 0.001, n= 8 for control and 11 for 0 Ca2+ and BAPTA pretreatment). B, plot of mean mEPSC frequency under normal ischaemia conditions (filled circles) and when BAPTA-AM presoaked slices are exposed to ischaemia with extracellular Ca2+ removed as in Fig. 5 (open circles). ANOVA indicated a significant effect of condition, P < 0.05, and ischaemia, P < 0.001, but no significant interaction, P > 0.8, indicating no block of increase in mEPSCs. The 0 Hz level is marked by the arrowhead. C, plot of mean ischaemia-induced increase in mEPSC frequency (expressed as a ratio to control mEPSC frequency for each condition) for slices under normal ischaemia conditions (filled circles) and for BAPTA-AM presoaked slices exposed to ischaemia with extracellular Ca2+ removed as in Fig. 5 (open circles). D, plot of mEPSC frequency in response to bath application of the Ca2+ ionophore ionomycin (12 μm, indicated by bar above plot) for control slices (filled circles) and for BAPTA-AM presoaked slices exposed to ionomycin with extracellular Ca2+ removed (open circles). Asterisks indicate time points where the percentage increase is significantly different between conditions (P < 0.05, as assessed by Bonferroni's post hoc test on the two conditions which showed a significant interaction with ionomycin, P < 0.001, n= 3 for control and for BAPTA-AM + 0 Ca2+). The 0 Hz level is marked by the arrowhead. E, plot of mean ionomycin-induced increase in mEPSC frequency (expressed as a ratio to control mEPSC frequency for each condition) for slices under normal conditions (filled circles) and for BAPTA-AM presoaked slices exposed to ionomycin with extracellular Ca2+ removed as in Fig. 5 (open circles).

Download figure to PowerPoint

As a positive control for the ability of our BAPTA-AM protocol to inhibit Ca2+-induced mEPSCs, we used a Ca2+ ionophore, ionomycin, to directly raise presynaptic [Ca2+]c (Fig. 7D and E). Bath application of ionomycin (12 μm) caused an increase in mEPSC frequency similar to that induced by ischaemia (Fig. 7D and E). However, in contrast to ischaemia, ionomycin had no effect on mEPSC frequency in the absence of external Ca2+, in slices that were pre-incubated in BAPTA-AM (Fig. 7D and E). Thus, our BAPTA-AM protocol does successfully prevent increases in mEPSCs triggered by rises in presynaptic [Ca2+]c. Taken together, our data indicate that the ischaemia-induced increase in mEPSC frequency is not triggered by intracellular Ca2+ from either influx or release from intracellular stores.

Ischaemia-induced vesicle release is mediated by actin filament depolymerization

Recent studies indicate that transient episodes of ischaemia cause actin filament depolymerization in dendritic spines, resulting in spine retraction that contributes to ischaemic preconditioning (Gisselsson et al. 2005; Meller et al. 2008). Since depolymerizing actin filaments in presynaptic terminals with latrunculin A causes Ca2+-independent vesicle release (Morales et al. 2000), we reasoned that ischaemia might increase vesicle release by depolymerizing actin filaments in presynaptic terminals, as it does in dendritic spines (Gisselsson et al. 2005; Meller et al. 2008). To test this hypothesis, we pre-treated slices with the actin filament stabilizer jaspamide (20 μm), which prevents the vesicle release induced by latrunculin A (Morales et al. 2000). In agreement with our hypothesis, jaspamide pretreatment prevented the ischaemia-induced increase in mEPSC frequency (Fig. 8A; mean peak mEPSC frequency in ischaemia was 0.30 ± 0.11 Hz, n= 9, for jaspamide soaked slices with a significant effect of condition, P < 0.001, and a significant interaction with ischaemia, P < 0.001, between the two conditions, with the points marked with asterisks being significantly different, P < 0.05). Addition of jaspamide to the recording pipette solution did not affect the ischaemia-induced increase in mEPSCs frequency (Fig. 8B; no significant affect of condition or interaction with ischaemia, P > 0.05) indicating that the effect of soaking slices in jaspamide was mediated by stabilizing actin filaments in presynaptic terminals. Pre-treating slices with another membrane-permeant actin filament stabilizer, phalloidin oleate (20 μm), also significantly reduced the ischaemia-induced increase in mEPSC frequency (Fig. 8B; mean peak mEPSC frequency in ischaemia was 0.60 ± 0.30 Hz, n= 5 for phalloidin soaked slices with a significant effect of ischaemia, P < 0.001, but a significant interaction with ischaemia, P < 0.0001, between the two conditions, with the points marked with asterisks being significantly different, P < 0.05).

image

Figure 8. Preventing actin filament depolymerization in presynaptic terminals blocks the ischaemia-induced increase in mEPSC frequency A, graph depicting mEPSC frequency change induced by ischaemia in control slices (black filled circles), slices pre-soaked in jaspamide (20 μm, grey filled circles), and in control slices but with jaspamide (20 μm) included in the recording pipette (open circles). Asterisks indicate time points where the percentage increase is significantly different between control and jaspamide soaked slices (P < 0.05, as assessed by Bonferroni's post hoc test on the two conditions which showed a significant interaction with ischaemia, P < 0.001, n= 16 for control and 9 for jaspamide soaked slices). ANOVA comparison of control versus jaspamide in the pipette showed no significant effect of condition nor significant interaction with ischaemia (P > 0.05, n= 5 for jaspamide in pipette). The 0 Hz level is marked by the arrowhead. B, graph depicting mEPSC frequency change induced by ischaemia in control slices (filled circles), and in slices pre-soaked in phalloidin oleate (20 μm, open circles). Asterisks indicate time points where the percentage increase is significantly different between control and phalloidin soaked slices (P < 0.05, as assessed by Bonferroni's post hoc test on the two conditions which showed a significant interaction with ischaemia, P < 0.0001, n= 26 for control and 5 for phalloidin soaked slices). The 0 Hz level is marked by the arrowhead. C, plot of mean percentage change in Fura-2 emission ratio under normal ischaemia conditions (filled circles) and in jaspamide pre-soaked slices (open circles). ANOVA indicated no significant difference between conditions or interactions with ischaemia, P > 0.05, n= 8 for control and 9 for jaspamide.

Download figure to PowerPoint

Jaspamide pre-treatment did not affect the ischaemia-induced increase in [Ca2+]c (Fig. 8C), further confirming the independence of the ischaemia-induced increase in mEPSCs from the rise in [Ca2+]c. The data suggest that the ischaemia-induced, increase in mEPSCs is due to actin depolymerization in presynaptic terminals, resulting in increased vesicle fusion.

The ischaemia-induced early rise in [Ca2+]c is not mediated by glutamate receptors

Although our data indicate that the increase in mEPSC frequency is not caused by the early rise in [Ca2+]c, given the permeability of glutamate receptors to Ca2+ and the parallel temporal relationship between mEPSC frequency and [Ca2+]c, it remains possible that the increase in glutamatergic mEPSCs causes the rise in [Ca2+]c. However, blocking ionotropic glutamate receptors alone or in conjunction with blocking metabotropic glutamate receptors, using the same pharmacological agents as shown in Fig. 4, did not prevent the ischaemia-induced early rise in [Ca2+]c (Fig. 9A; mean peak increase of Fura-2 emission ratio was 31.57 ± 0.04%, n= 4, and 44.14 ± 8.56%, n= 4, for d-AP5+NBQX and d-AP5+NBQX+(S)-MCPG, neither different from control ischaemia, nor showing a significant interaction, P > 0.05). To rule out the possibility that a small effect of glutamate receptor antagonists was obscured by interslice variability, we also applied the glutamate receptor antagonists acutely during the ischaemia-induced rise in [Ca2+]c, also without any observable effect (Fig. 9B and C; no significant affect of condition or interaction with ischaemia, P > 0.05, n= 3). Thus, despite their temporal coincidence, the ischaemia-induced increase in mEPSCs and [Ca2+]c are independent of each other.

image

Figure 9. Blocking glutamate receptors does not affect the early ischaemia-induced rise in [Ca2+]c A, plot of mean percentage change in Fura-2 emission ratio during ischaemia alone (black filled circles), in the presence of d-AP5 (50 μm) and NBQX (25 μm, grey filled circles), and in the presence of d-AP5 (50 μm), NBQX (25 μm) and (S)-MCPG (500 μm, open circles). There were no significant differences between conditions or interactions with ischaemia (P > 0.05, n= 8, 4, and 4, respectively). B, plot of percentage change in Fura-2 emission ratio for a single representative slice during ischaemia alone and upon going into ischaemia supplemented with d-AP5 (50 μm), NBQX (25 μm) and (S)-MCPG (500 μm), both indicated by bars above trace. Note there was no obvious deflection in Fura-2 emission upon going into glutamate receptor blockers. C, plot of mean percentage change in Fura-2 emission ratio for slices during ischaemia alone (filled circles) and for slices that during ischaemia were exposed to d-AP5 (50 μm), NBQX (25 μm) and (S)-MCPG (500 μm, open circles). The bar above traces indicating the time of glutamate blocker exposure applies only to the open circles; the filled circles are for the control ischaemia slices. In addition to the lack of any obvious deflection in Fura-2 emission upon going into glutamate receptor blockers, ANOVA indicated no significant differences between conditions or interactions with ischaemia (P > 0.05, n= 8 for control and 3 for blockers applied during ischaemia).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix

The principle findings of this study are that simulated ischaemia causes an early rise in [Ca2+]c that parallels an early increase in mEPSC frequency, both of which precede the ischaemic depolarization (ID) by several minutes. The ischaemia-induced rise in [Ca2+]c is mediated in part by Ca2+ influx across the plasma membrane and in part from release from intracellular stores. However, the increase in [Ca2+]c and mEPSC frequency are independent of each other as preventing the increase in [Ca2+]c does not affect the increase in mEPSC frequency, and blocking the glutamate receptors that underlie the mEPSCs does not affect the rise in [Ca2+]c. In contrast, the ischaemia-induced increase in mEPSCs was prevented by blocking actin filament depolymerization, indicating that the vesicle release is triggered by actin filament depolymerization in presynaptic terminals.

Our observation of an early increase in mEPSC frequency confirms previous reports examining anoxia/ischaemia in a variety of brain slice preparations (Katchman & Hershkowitz, 1993; Fleidervish et al. 2001; Allen et al. 2004). Our observation of an early increase in [Ca2+]c agrees with a previous study of ischaemic hippocampus in vivo (Silver & Erecinska, 1990), but conflicts with a previous study of simulated ischaemia in hippocampal slices (Mitani et al. 1993). The apparent contradiction may reflect the different affinities of the Ca2+-sensitive fluorophores used (rhod-2 in the latter study and Fura-2 in our own), and suggests that the early rise in [Ca2+]c is relatively modest, requiring a high affinity probe such as Fura-2 to be detected. Alternatively, the difference may be explained by the different methods for simulating ischaemia. Mitani et al. (1993) simply removed oxygen and glucose (oxygen and glucose deprivation, OGD), whereas we supplemented OGD with iodoacetic acid (IAA) to block glycolysis. If this methodological difference underlies the different Ca2+ responses in the respective slice studies, then our study's agreement with the in vivo study (Silver & Erecinska, 1990) suggests that for simulating ischaemia in vitro, supplementing OGD with IAA, which is presumably more severe than OGD alone, more closely replicates the conditions of ischaemia in vivo.

A previous study with hippocampal slices showed that the ryanodine receptor antagonist dantrolene prevented the anoxia-induced increase in mEPSC frequency (Katchman & Hershkowitz, 1993). Based on that finding, the authors suggested that Ca2+ release from intracellular stores in presynaptic terminals triggered the early increase in mEPSCs. While we do observe an early release of Ca2+ from intracellular stores that parallels the increase in mEPSC frequency (Fig. 5), this calcium is unlikely to be the trigger for the increase in mEPSCs, under our conditions, since it can be blocked without preventing the ischaemia-induced increase in mEPSCs (Fig. 7). It is possible that there are simply different triggers during anoxia and ischaemia, with ischaemia likely to be leading to a more rapid decline in ATP levels. An alternative explanation is that dantrolene interferes with anoxia-induced vesicle release either indirectly or via mechanisms other than blocking the ryanodine receptor (Krnjevic & Xu, 1996; Salinska et al. 2008).

Source of early rise in [Ca2+]c

Our imaging protocol detects [Ca2+]c signals in multiple subcellular compartments (Fig. 4), making it difficult to clearly define which compartment(s) lead to the ischaemia-induced increase in [Ca2+]c we have observed. However, in vivo studies of ischaemia have shown that hippocampal pyramidal cells exhibit an early rise in [Ca2+]c, similar in magnitude and time course to the rise that we observe in our bulk loaded slices (Silver & Erecinska, 1990). Furthermore, since the relationship between mEPSC frequency and presynaptic [Ca2+]c is steep above the resting [Ca2+]c (Frerking et al. 1997) but shallow below it, if ischaemia caused early changes in presynaptic [Ca2+]c we would expect it to affect mEPSC frequency. Moreover, preventing those changes by buffering Ca2+ should reduce the impact of ischaemia on mEPSC frequency, which we did not observe (Fig. 7). Accordingly, while the particular subcellular compartment in which the rise in [Ca2+]c occurs remains obscure, we think it unlikely that this increase takes place in presynaptic terminals.

Removal of external Ca2+ reduced the ischaemia-induced early rise in [Ca2+]c (Fig. 5), demonstrating that some of the rise is due to Ca2+ influx across the plasma membrane. This is in agreement with in vivo ischaemia studies showing that the early rise in [Ca2+]c in pyramidal cells is accompanied by a parallel decrease in [Ca2+] in the extracellular space (Silver & Erecinska, 1990). There are many potential mechanisms by which Ca2+ might enter pyramidal cells or astrocytes during ischaemia (Duffy & MacVicar, 1996; Lipton, 1999; Zhang & Lipton, 1999; Xiong et al. 2004; Bondarenko et al. 2005; Thompson et al. 2006; Rossi et al. 2007), but we can definitively rule out glutamate-gated channels for the early rise in [Ca2+]c because blocking these channels did not affect the rise in [Ca2+]c (Fig. 9). The fact that the early rise in [Ca2+]c was not accompanied by a detectable membrane current in voltage-clamped pyramidal cells also argues against a role for voltage-gated channels, or for channels with a large unitary conductance such as hemi-gap junctions (Thompson et al. 2006). Silver & Erecinska (1990) suggested that the early influx of Ca2+ was mediated by the Na+–Ca2+ exchanger pumping out Na+ ions that entered via the Na+–H+ exchanger. While the Na+–Ca2+ exchanger is electrogenic (Hinata & Kimura, 2004), its slow rate of transport compared to channels would generate much smaller currents that would be difficult for us to detect.

Ischaemia induces Ca2+-independent vesicle release due to actin filament depolymerization in presynaptic terminals

Our data demonstrate that the early phase of ischaemia induces vesicle release independently of changes in [Ca2+]c (Fig. 7). Even if there was a small rise in [Ca2+]c in presynaptic terminals that was not detected by our recording methodology, the available evidence suggests that the typical resting [Ca2+]c (∼100 nm) is below the steep phase of the relationship between [Ca2+]c and quantal vesicle release rate (Frerking et al. 1997; Xu et al. 2009). Therefore, since our BAPTA loading protocol reduced the resting [Ca2+]c (Fig. 5C) and reduced, if not abolished, the ischaemia-induced rise in [Ca2+]c (Fig. 7A), any increase in mEPSCs that was triggered by an undetected rise in [Ca2+]c should be proportionally smaller, as we observed experimentally for ionomycin-induced mEPSCs (Fig. 7E). However, we observed a greater proportional increase in mEPSC frequency during ischaemia in BAPTA loaded slices (Fig. 7C), suggesting the recruitment of a Ca2+-independent mechanism.

Both hypertonic sucrose and α-latrotoxin can also induce vesicle release independent of changes in presynaptic [Ca2+]c (Khvotchev et al. 2000), but the underlying mechanisms are not well understood. The Ca2+-independent actions of α-latrotoxin probably involve binding to latrophilin (Volynski et al. 2003), but we are not aware of any reports regarding latrophilin and ischaemia. In contrast, latrunculin A has a well-understood effect as an actin depolymerizing agent, and is known to cause a Ca2+-independent release of vesicles, presumably by depolymerizing actin filaments that normally restrain vesicle fusion (Morales et al. 2000). Thus, an ischaemia-induced reorganization of the actin cytoskeleton might underlie the Ca2+-independent increase in mEPSCs observed here.

The previous literature is consistent with this idea, as early events of brain ischaemia include both depolymerization of actin filaments (Gisselsson et al. 2005) and a repositioning of glutamatergic vesicles toward the plasma membrane (Williams & Grossman, 1970). Moreover, we have found that preventing actin filament depolymerization by presoaking slices in jaspamide, an actin filament stabilizing agent which prevents vesicle release triggered by latrunculin A (Morales et al. 2000), abolished the ischaemia-induced increase in mEPSCs (Fig. 8A), as did another actin stabilizing agent, phalloidin oleate (Fig. 8B). Inclusion of jaspamide in the recording pipette did not block the ischaemia-induced increase in mEPSCs (Fig. 8A) suggesting that the block by pretreating slices was mediated by stabilizing actin filaments in presynaptic terminals, rather than in postsynaptic spines. Thus, the early increase in mEPSC frequency during ischaemia is due to actin filament depolymerization in presynaptic terminals and a corresponding release of vesicles, similar to that which occurs in response to latrunculin A.

Ischaemia-induced increases in mEPSC frequency may trigger ischaemic preconditioning

Either transient episodes of ischaemia or exogenously applied glutamate analogues can trigger actin depolymerization-mediated postsynaptic spine retraction (Hasbani et al. 2001), and a resultant protected state (Halpain et al. 1998; Hasbani et al. 2001; Ikegaya et al. 2001; Graber et al. 2004). Furthermore, at least in some preparations, glutamate receptor antagonists can prevent the ischaemia-induced spine retraction (Jourdain et al. 2002). These observations suggest that a sub-toxic release of glutamate during the early stages of ischaemia can be a trigger for spine retraction and ischaemic preconditioning. The increase in glutamatergic vesicle release reported here and elsewhere is an obvious candidate trigger for this process.

Given the role of glutamate receptor activation in delayed cell death (Simon et al. 1984; Gill et al. 1987; Pivovarova et al. 2004), it is possible that the ischaemia-induced increase in mEPSCs could trigger delayed cell death rather than preconditioning. However, for CA1 pyramidal cells, delayed cell death is dependent on Ca2+ influx through NMDA receptors (Pivovarova et al. 2004), and in our study blocking glutamate receptors did not affect the ischaemia-induced elevation in [Ca2+]c (Fig. 9). Presumably, the reason the ischaemia-induced increase in mEPSCs does not affect [Ca2+]c is because cells remain hyperpolarized, which due to voltage-dependent Mg2+ block, prevents Ca2+ influx through NMDA receptor channels (Mayer et al. 1984). Furthermore, in hippocampal slices, removing extracellular Ca2+ can prevent glutamate receptor triggered cell death without preventing changes in dendritic morphology or preconditioning, the latter effects being triggered by Na+ influx (Ikegaya et al. 2001). Accordingly, we favour our preconditioning trigger hypothesis. Thus, actin filament depolymerization may play a dual role in ischaemic preconditioning, with presynaptic depolymerization driving vesicular glutamate release, which then triggers glutamate-induced depolymerization postsynaptically to induce spine retraction and subsequent protection.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix

Appendix

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix

Author contributions

Both of the authors have contributed to conception and design of the experiments, collection, analysis and interpretation of data, and drafting the article or revising it critically for important intellectual content. The studies were carried out at the Department of Behavioral Neuroscience, Oregon Health and Science University. All the authors approved the final version to be published.

Acknowledgements

We would like to acknowledge Drs Matt Frerking, Henrique von Gersdorff and Jun Hee Kim for helpful discussions about experimental design. This study was supported by National Institute of Neurological Disorders and Stroke Grant R01NS051561. The authors disclose no conflicts.