Intracellular Zn²⁺ accumulation enhances suppression of synaptic activity following spreading depolarization

Abstract

Spreading depolarization (SD) is a feed-forward wave that propagates slowly throughout brain tissue and recovery from SD involves substantial metabolic demand. Presynaptic Zn²⁺ release and intracellular accumulation occurs with SD, and elevated intracellular Zn²⁺ ([Zn²⁺]_i) can impair cellular metabolism through multiple pathways. We tested here whether increased [Zn²⁺]_i could exacerbate the metabolic challenge of SD, induced by KCl, and delay recovery in acute murine hippocampal slices. [Zn²⁺]_i loading prior to SD, by transient ZnCl₂ application with the Zn²⁺ ionophore pyrithione (Zn/Pyr), delayed recovery of field excitatory post-synaptic potentials (fEPSPs) in a concentration-dependent manner, prolonged DC shifts, and significantly increased extracellular adenosine accumulation. These effects could be due to metabolic inhibition, occurring downstream of pyruvate utilization. Prolonged [Zn²⁺]_i accumulation prior to SD was required for effects on fEPSP recovery and consistent with this, endogenous synaptic Zn²⁺ release during SD propagation did not delay recovery from SD. The effects of exogenous [Zn²⁺]_i loading were also lost in slices preconditioned with repetitive SDs, implying a rapid adaptation. Together, these results suggest that [Zn²⁺]_i loading prior to SD can provide significant additional challenge to brain tissue, and could contribute to deleterious effects of [Zn²⁺]_i accumulation in a range of brain injury models.

INTRODUCTION

Spreading depolarizations (SD) are waves of neuronal and glial depolarization that propagate slowly throughout brain tissue (Leao 1944, Somjen 2001). SD is commonly recorded as a large extracellular negative direct current (DC) potential shift and results in a transient suppression of both electrocorticographic activity (Leao 1947, Somjen 2001). Large amounts of energy are required to restore ionic gradients before normal function can return, as demonstrated by large reductions in ATP and glucose levels following SD (Shinohara et al. 1979, Mies & Paschen 1984). Extracellular adenosine accumulation occurs with severe metabolic challenges (Dale et al. 2000, Dunwiddie & Masino 2001), and can contribute to the transient suppression of field excitatory postsynaptic potentials (fEPSPs) in brain slice models of SD (Lindquist & Shuttleworth 2012). In healthy brain tissue SD is not injurious (Nedergaard & Hansen 1988), however under conditions where metabolism is already compromised, such as in ischemic brain tissue, the additional metabolic challenge of repetitive SD events can lead to irrecoverable damage (Hartings et al. 2003, Busch et al. 1996). A growing body of literature from both animal and human recordings suggests that clusters of SDs can be a significant contributing factor to the enlargement of acute brain injuries (Dreier 2011, Hartings et al. 2011a, Lauritzen et al. 2011, Oliveira-Ferreira et al. 2010, Nakamura et al. 2010).

We have recently demonstrated a significant release of synaptic Zn²⁺ following SD, and subsequent intracellular accumulation (Carter et al. 2011). Zn²⁺ is highly regulated and normally kept at very low intracellular levels, however excessive intracellular Zn²⁺ ([Zn²⁺]_i) accumulation has been associated with a wide range of brain injuries including ischemia and trauma (Shuttleworth & Weiss 2011, Sensi et al. 2011). It is not yet known whether release and accumulation of Zn²⁺ following SD may be a significant contributor to Zn²⁺ toxicity in a variety of brain injury settings. One mechanism for toxic effects of Zn²⁺ involves disruption of cellular energy production. Zn²⁺ can directly bind to and inhibit both glyceraldehyde-3-phosphate dehydrogenase (GAPDH, (Sheline et al. 2000)) and the α-ketoglutarate dehydrogenase complex (KGDHC, (Brown et al. 2000)), inhibiting glycolysis and the TCA cycle, respectively. Additionally, Zn²⁺ can inhibit several steps in the mitochondrial electron transport chain (ETC) (Dineley et al. 2003, Sharpley & Hirst 2006). Indirectly, Zn²⁺ can also inhibit metabolism by activating NAD⁺-catabolizing enzymes such as poly(ADP-ribose) polymerase 1 (PARP-1) and the sirtuin family of proteins (Cai et al. 2006, Sheline et al. 2000), depleting NAD⁺ levels and inhibiting glycolysis.

The aim of the current study was to determine whether increasing [Zn²⁺]_i can exacerbate the delay in recovery from SD. The results suggest that when [Zn²⁺]_i levels are sufficiently elevated prior to the passage of SD, there is a marked additional metabolic challenge, leading to delayed recovery of synaptic transmission.

MATERIALS AND METHODS

Slice preparation and recording

All procedures were carried out in accordance with the National Institutes of Health guidelines for the humane treatment of laboratory animals, and the protocol for these procedures was reviewed annually by the Institutional Animal Care and Use Committee (IACUC) at the University of New Mexico School of Medicine. Acute hippocampal brain slices (350µm) were prepared from wild type C57Bl/6J and ZnT3 KO mice of either sex (4–8 weeks old). After cutting in ice-cold cutting solution and then holding for 1hr at 35°C, artificial cerebrospinal fluid (ACSF) was changed, and slices were held at room temperature until used for recording. Individual slices were then transferred to the recording chamber, and superfused with oxygenated ACSF at 2mL/min at 32°C. Mice homozygous for deletion of the ZnT3 gene were originally obtained from Dr. Richard Palmiter (Cole et al. 1999) and backcrossed onto C57Bl/6J for >13 generations (LSUHSC). A homozygous breeding colony of ZnT3 KO animals was then established at the University of New Mexico.

Extracellular measurements of field excitatory postsynaptic potentials (fEPSPs), were made via stimulation of Schaffer collaterals with a concentric bipolar electrode (CBAEC75, FHC, Bowdoin, ME, USA). Stimulation intensity was set to achieve ~60% maximum response during recording sessions. Extracellular recordings of fEPSPSs and direct current (DC) shifts were made by glass microelectrodes (~2–5MΩ), filled with ACSF and placed in stratum radiatum of CA1 ~50 µm below the surface of the slice. Spreading depolarization (SD) was induced by brief pressure pulses (30psi, 80–100ms) of 1M KCl from a homemade picospritzer (UniBlitz Model D122, Vincent Associates, Rochester, NY, USA; valve: 3-165-900, General Valve Corporation, Fairfield, NY, USA). Onset and propagation of SD was monitored by intrinsic optical signals (IOS), generated by transmission of red light through the tissue (>575nm). Propagation rate and IOS analyses were performed in TillVision (Till Photonics, Rochester, NY, USA). Comparison of fEPSP recovery between groups was defined by the time for fEPSPs to return to 50% pre-SD levels (t₅₀), unless stated otherwise.

Intracellular Zn²⁺ loading

To simulate conditions in which [Zn²⁺]_i levels were increased, we used a transient exposure (20min) to pyrithione (Pyr, 5µM) in the presence of 0, 30, or 100µM ZnCl₂ (Zn/Pyr). This was followed by a 10min washout prior to SD induction to minimize possible contributions of increased extracellular Zn²⁺ during the passage of SD (Aiba et al. 2012). Histidine (His, 200µM) was present during all Zn/Pyr exposures to help solubilize Zn²⁺ and prevent it from precipitating out of solution (Rumschik et al. 2009).

Adenosine measurements

Measurements of extracellular adenosine accumulation were performed as previously described (Lindquist & Shuttleworth 2012). Briefly, enzyme-coupled electrochemical probes (SBS-ADO-05-50, Sarissa Biomedical, Coventry, UK) generate an amperometric signal proportional to the concentration of adenosine. Measurements reported here are estimates of accumulation at the slice surface, and synaptic levels may be different due to transporter activities within the slice. Calibration of the probes with exogenous adenosine (5–50µM) was performed immediately following individual recording sessions. It is noted that in the present study calibration stocks were made daily, and estimates of adenosine levels following SD are ~10-fold lower than previously reported (Lindquist & Shuttleworth 2012). Adenosine probes were positioned between the bipolar stimulating and recording electrodes in stratum radiatum of CA1.

Reagents and solutions

Slice cutting solution contained (in mM): 3 KCl, 1.25 NaH₂PO₄, 6 MgSO₄, 26 NaHCO₃, 0.2 CaCl₂, 10 glucose, 220 sucrose, and 0.43 ketamine. ACSF contained (in mM): 126 NaCl, 3 KCl, 1.25 NaH₂PO₄, 1 MgSO₄, 26 NaHCO₃, 2 CaCl₂, and 10 glucose, equilibrated with 95%O₂ / 5%CO₂. N,N,N’,N’-tetrakis(2-pyridylmethyl)-ethylenediamine (TPEN) was obtained from Intvitrogen (Carlsbad, CA, USA). All other reagents were obtained from Sigma (St. Louis, MO, USA).

Statistical Analysis

Data are reported as mean±standard error of the mean (S.E.M.). Statistical analysis was performed in GraphPad Prism (La Jolla, CA, USA) using paired or unpaired Student’s t-test, or analysis of variance (ANOVA), with Bonferroni correction for multiple comparisons where appropriate. Differences were considered significant when p<.05.

RESULTS

1. Effects of transient [Zn²⁺]_i loading on fEPSP suppression following SD

Transient loading of [Zn²⁺]_i (using pyrithione, see methods) prior to SD resulted in significant delay in recovery of fEPSPs after SD. In slices not loaded with Zn²⁺ (5µM Pyr control), SD resulted in fEPSP amplitude suppression that took 11.22±0.95min (n=6) to return to baseline levels (Fig 1A), similar to previously published data (Lindquist & Shuttleworth 2012). [Zn²⁺]_i loading produced a concentration-dependent increase in the duration of fEPSP suppression following SD (Fig 1A) and final fEPSP amplitudes remained substantially depressed 25min following SD with the higher Zn²⁺ concentrations (30 & 100µM).

Figure 1. Transient [Zn²⁺]_i loading resulted in increased fEPSP suppression following SD.

A: Plots showing the effects of a single SD on fEPSP amplitude. SDs were generated at the time indicate (arrow), and resulted in a rapid and near-complete initial suppression of fEPSP amplitude. Under control conditions (Pyr), this was followed by recovery over the next ~11min to baseline levels. Slices treated with Zn²⁺ together with the ionphore (Zn/Pyr, applied during time indicated by the black bar), resulted in a concentration-dependent increase in secondary suppression following SD (p<.0001, ANOVA with Bonferroni post hoc test, 5µM Pyr vs. 10, 30, and 100µM Zn/Pyr, n=6 each). Circle: Pyr; Square: 30µM Zn/Pyr; Diamond: 100µM Zn/Pyr, n=6 each. B: Plots showing effects of Zn/Pyr exposure on fEPSP amplitudes in the absence of SD in a separate set of slices (Circle: Pyr; Square: 30µM Zn/Pyr; Diamond: 100µM Zn/Pyr, n=6 each). C: Representative fEPSPs from time points indicated in (A): a-d, and (B): e-h. D: Mean data showing final fEPSP amplitudes (measured at 60min), determined from experiments shown in (A) & (B) (* p<.01unpaired Student’s t-test, n=6 each).

We noted a slight run-down of fEPSP amplitude immediately prior to the onset of SD when slices were loaded with Zn²⁺. Therefore a set of time-control studies was completed to test whether progressive run-down of fEPSP amplitude alone (without SD) could have produced the results seen. Thus slices were loaded with 0, 30, or 100µM Zn/Pyr (20min followed by wash out) without inducing SD (Fig 1B). Figures 1C&D compare the effects of [Zn²⁺]_i loading on final fEPSP amplitudes (assessed at the same time point, i.e. 60min from start of recording) with and without the additional challenge of SD. A concentration-dependent increase in fEPSP run-down was detected even without SD, but the degree of suppression was greatly increased by the combination of Zn²⁺ with SD.

2. [Zn²⁺]_i loading increased DC shift duration

We next assessed the characteristics of SDs, generated under the conditions shown in Figure 1A. In Pyr control slices, the waveform and duration of DC shifts were similar to those previously described (Somjen 2001) and recorded under near-identical conditions without Pyr (Aiba & Shuttleworth 2012) (Figure 2A). Following [Zn²⁺]_i loading, the duration of DC shifts were markedly increased (Figure 2B) in a concentration-dependent manner. Propagation rates of SD were assessed using IOS signals (see Methods) and no significant changes in rates were observed (Figure 2C). The effects of [Zn²⁺]_i loading on DC shift duration are consistent with previous reports that impaired metabolism is associated with increased DC shift duration and worse clinical outcome (Hartings et al. 2011b).

Figure 2. [Zn²⁺]_i loading resulted in prolonged DC shifts.

A: Representative extracellular DC potential shifts recorded during SD in control conditions (Pyr) or 30 and 100µM Zn/Pyr. B: Mean data from recordings as shown in (A) showing a concentration-dependent increase in DC duration (measured at 80% recovery, * p<.0001, ANOVA with Bonferroni post hoc test, n=6 each). C: [Zn²⁺]_i loading had no significant effect on SD propagation rate, as measured from intrinsic optical signals in the same data set as shown in (B) (p=.15, ANOVA with Bonferroni post hoc test, 5µM Pyr vs. 30 and 100µM Zn/Pyr, n=6 each).

3. Extracellular adenosine accumulation and A1R activation following SD with [Zn²⁺]_i loading

Impaired metabolism can lead to increased adenosine accumulation (Dunwiddie & Masino 2001). It was recently shown in brain slices that secondary fEPSP suppression after SD is mainly mediated by extracellular adenosine accumulation and activation of adenosine A1 receptors (A1Rs) (Lindquist & Shuttleworth 2012). As shown in Figure 3A, pre-exposure to a selective A1R antagonist (8-cyclopentyl-1,3-dipropylxanthine, DPCPX, 200nM), significantly decreased the duration and degree of fEPSP suppression after SD with [Zn²⁺]_i loading, compared to the suppression shown in Figure 1. The degree of fEPSP suppression under these different conditions is summarized in Figure 3B. These results are consistent with the effective reversal of secondary fEPSP suppression described previously (Lindquist & Shuttleworth 2012), and imply that the sustained suppression observed in Figure 1A is largely due to A1 receptor activation rather than neuronal injury.

Figure 3. A1R activation and enhanced adenosine accumulation after SD with [Zn]_i loading.

A: Experiments conducted in the presence of an adenosine A1R antagonist, DPCPX (200nM) showed substantially larger recovery of fEPSPs following SD (arrow) after Zn/Pyr exposure (black bar) (compare with Figure 1, n=6 each). B: Comparison of fEPSP suppression under control and DPCPX conditions. Suppression of fEPSP for each preparation was calculated as the area between baseline amplitude and the recovered fEPSP amplitude, from immediately following the onset of SD to the end of the recording (25 min total duration). Suppression was significantly decreased by DPCPX under all conditions tested (*p<.01, **p<.001 unpaired Student’s t-test, n=6 each). C: Amperometric measurements using an adenosine-sensitive electrochemical probe, showing responses to SD as compared with calibration exposures to exogenous adenosine (5µM) to generate estimated adenosine concentrations at the slice surface (see Methods). Adenosine accumulation during SD (arrow) was significantly enhanced by prior exposure to 30µM Zn/Pyr, compared with Pyr controls (p<.001 at peak, n=5 each). D: Application of exogenous adenosine immediately following SD induction (arrow) resulted in a significant reduction in fEPSP recovery that mimicked the effect of Zn/Pyr exposure (n=5 each).

We next tested whether additional extracellular adenosine accumulation contributed to the A1R-mediated fEPSP suppression seen above. Estimated extracellular adenosine concentration was measured at the slice surface using an enzyme linked probe (see Methods). A test concentration of 30µM Zn/Pyr in combination with SD was chosen as these conditions resulted in a substantial suppression of fEPSPs (see Fig 1A) that was completely mediated by A1R activation (see Fig 3A). Estimated peak extracellular adenosine accumulation during SD was significantly greater when [Zn²⁺]_i loading was combined with SD (Fig 3C, n=5 each). To test whether exogenous adenosine could replicate the effects of Zn/Pyr exposure, adenosine (50µM) was applied immediately following the induction of SD in a separate set of experiments (Fig. 3D). The recovery of fEPSPs following SD was significantly decreased by exogenous adenosine, compared with control slices (Fig. 3D). Together, these findings suggest that the enhanced fEPSP suppression by [Zn²⁺]_i loading with SD was mediated predominantly by increased extracellular adenosine and A1R activation.

4. Potential intracellular targets for [Zn²⁺]_i loading effects

We next tested potential mechanisms which could link [Zn²⁺]_i loading to metabolic inhibition. Enhanced levels of [Zn²⁺]_i can deplete NAD⁺ and thereby inhibit glycolysis either by induction of poly(ADP-ribose) polymerase 1 (PARP-1) or activation of SIRT protein pathways (Sheline et al. 2000, Cai et al. 2006). These possibilities were tested by using selective inhibitors for either PARP-1 (3,4-dihydro-5-[4-(1-piperidinyl)butoxyl]-1(2H)-isoquinolinone DPQ, 25µM) or SIRT (sirtinol, 30µM). As shown in Figure 4, DPQ with 30µM Zn/Pyr did not increase fEPSP recovery compared to vehicle treated slices with 30µM Zn/Pyr (p=.073, n=8 each). In addition, slices pre-exposed to sirtinol with 30µM Zn/Pyr did not result in a significant recovery of fEPSPs following SD when compared with preparations exposed to Zn/Pyr alone (p=.8, n=3, data not shown).

Figure 4. Lack of effect of PARP-1 inhibition or pyruvate supplementation on [Zn²⁺]_i loading effects.

A: Slices exposed to the PARP-1 inhibitor DPQ (25µM) showed little difference in recovery of fEPSP following SD after 30µM Zn/Pyr compared to slices exposed to 30µM Zn/Pyr alone (p=.13, n=8 each). DMSO (0.1%) was used as a vehicle in these studies. B: Addition of 5mM pyruvate together with 30µM Zn/Pyr had no effect on the recovery of fEPSPs following SD compared to slices exposed to 30µM Zn/Pyr alone (p=.60, n=6 each).

In a third set of slices, 5mM pyruvate was superfused for the entire recording session to provide additional TCA metabolic supplementation. In the presence of 5mM pyruvate, fEPSPs suppression following SD was not significantly different compared to slices exposed to 30µM Zn/Pyr alone (Fig 4B, n=6 each). Additionally, in slices supplemented with lactate (5mM), fEPSP suppression following 30µM Zn/Pyr with SD was similar to slices without exogenous lactate (p=.6, n=3, data not shown). Taken together, these results suggest that [Zn²⁺]_i loading with SD may lead to increased extracellular adenosine accumulation via mechanisms downstream of glycolysis and pyruvate utilization.

5. Intracellular accumulation prior to SD is required for Zn²⁺ effects

To test whether the enhanced fEPSP suppression (seen in Fig 1A) was indeed a result of [Zn²⁺]_i accumulation and not a result of an extracellular of Zn²⁺, a set of slices was exposed to 30µM ZnCl₂ in the absence of Pyr. Figure 5A shows full recovery of fEPSPs to pre-SD levels following 30µM ZnCl₂ application (n=6), indicating that the enhanced suppression of fEPSPs seen in Figure 1A was due to intracellular accumulation and action of Zn²⁺. As seen in Figure 1, a small but significant decrease in fEPSP amplitude was also seen upon the addition of 30µM ZnCl₂ prior to SD induction (baseline: 101.2±1.62 vs. pre-SD: 89.41±3.05, p<.01).

Figure 5. Intracellular accumulation and exposure time needed for [Zn²⁺]_i loading effects.

A: Slices exposed to extracellular ZnCl₂ alone (black bar, without the ionophore to enhance intracellular accumulation) did not result in slowed recovery of fEPSPs following SD (compare with suppression in Figure 1) (n=6). Representative fEPSPs are shown at times (a-d) indicated on main plot. B: Similarly, no significant delay in fEPSP recovery was observed if [Zn²⁺]_i loading (Zn/Pyr, black bar) was applied beginning only 5min prior to SD induction (n=6). Representative fEPSPs are shown at times (a-d) indicated on main plot.

We next tested whether [Zn²⁺]_i loading prior to SD was required for the observed effects on fEPSPs (as seen in Fig 1A). To test this, Zn/Pyr application was begun only 5min before SD induction to allow equilibration within the slice prior to the SD challenge. In contrast to the results shown in Figure 1, this exposure resulted in no combined effect of 30µM Zn/Pyr with SD (Fig 5B). Together, these data suggest sufficient time is required (>5min) to allow for [Zn²⁺]_i accumulation and activation of intracellular targets to generate a significant effect on fEPSP suppression when combined with SD.

6. Contributions of endogenous Zn²⁺

We next evaluated whether intracellular accumulation of endogenous Zn²⁺ contributes to the metabolic burden of SD, and prolonged fEPSPs recovery. We first tested whether removal of all chelatable Zn²⁺ with a membrane permeable Zn²⁺ chelator (TPEN, 20µM throughout entire recording session) would influence fEPSP recovery from SD. TPEN pre-exposure resulted in a significant increase in the rate of fEPSP recovery after a single SD challenge (Fig 6A). Interestingly, there also was a significant increase in the fEPSP amplitude following SD in TPEN that was not observed in vehicle slices (TPEN baseline vs. post-SD amplitude: 99.80±1.14 vs. 107.6±1.40, p<.01; Vehicle baseline vs. post-SD amplitude: 99.32±0.74 vs. 100.5±1.99, p=.58, n=6 all groups).

Figure 6. Contributions of endogenous Zn²⁺.

A: A small but significant effect of the membrane permeable Zn²⁺ chelator TPEN on fEPSP recovery (20µM TPEN in 0.1% DMSO, 15 min pre-exposure and maintained throughout) (p=.04 for 50% recovery time, Veh vs. TPEN, n=6 each). B: Responses to three consecutive SD events in the same slices as (A) (20min intervals, onset indicated by arrows). TPEN exposure (20µM, 15 min pre-exposure and maintained throughout) revealed no cumulative effects due to progressive [Zn²⁺]_i accumulation on fEPSP recovery rate after repetitive SDs. C&D: Lack of effect of synaptically-released Zn²⁺ on fEPSP recovery rate after both single SDs (C) and repetitive SDs (D). Experiments are as described for A&B, but using comparison of wild type and ZnT3 KO slices rather than TPEN. p=.71 for 50% recovery time in C, and n=6 for all traces.

We next tested potential sources of endogenous Zn²⁺ responsible for these effects. Zn²⁺ can be accumulated to high concentrations in many glutamatergic terminals, as a consequence of loading via the transporter ZnT3 (Cole et al. 1999). As terminals depolarize during the propagation of SD, a significant release of synaptic Zn²⁺ occurs and can be visualized as a propagating wave of extracellular Zn²⁺ accumulation (Carter et al. 2011). Since significant postsynaptic [Zn²⁺]_i accumulation also occurs after SD (Carter et al., 2011), we investigated here whether endogenous synaptic Zn²⁺ release during SD could provide an additional metabolic burden to subsequent SDs. To test this, a series of three SDs were induced at 20min intervals, to allow sufficient time for uptake and activation of intracellular targets following Zn²⁺ release from the first SD (see section 5 above) to influence recovery from prior SDs. However, as shown in Figure 6B, there was no stepwise increase in the duration of fEPSP recovery after multiple SDs, suggesting that synaptic Zn²⁺ release may not be a significant contributor. Figures 6C&D are consistent with this, as there was no difference in fEPSP recovery after SD in slices from ZnT3 knockout animals, when compared with wild-type animals.

The results shown in Figure 6 were surprising, since we previously showed that there is no evidence of run-down of synaptic Zn²⁺ release during repetitive SDs (Carter et al., 2011), and the figures above suggest that repetitive [Zn²⁺]_i accumulation in advance of SD should lead to delayed functional recovery. Figure 7 provides an explanation for this apparent discrepancy, showing two consecutive SDs separated by 20min, with similar recovery rates as shown in Figure 6D. We next tested whether exogenous [Zn²⁺]_i loading with Zn/Pyr was still effective in suppressing fEPSPs in these slices following repetitive SD. Interestingly, the effects of Zn/Pyr on the enhanced fEPSP suppression described above (in SD-naïve slices, Fig 1A) were completely lost under conditions following repetitive SD (Fig 7A, n=6). A set of time-control experiments showed that exogenous [Zn²⁺]_i loading was still effective in slices where repetitive SDs were not generated (Fig 7B).

Figure 7. Loss of effectiveness of exogenous [Zn²⁺]_i loading, if slices were first subjected to repetitive SDs.

A: In slices exposed to 2 SDs before loading with 30µM Zn/Pyr exposure (black bar), the expected suppression of fEPSP following a subsequent (third) SD challenge was completely lost (n=6, compare with Fig 1A). Representative fEPSPs are shown above at times (a-d) indicated on main plot. B: Mean time control data, showing that exogenous [Zn²⁺]_i loading was still effective when given in SD-naïve slices, at the same time point as shown in (A) (n=6). Representative fEPSPs are shown above at times (a-d) indicated on main plot. Note the difference at time point d compared to the same time point in (A). C: Estimated extracellular adenosine accumulation at the surface of slices during SD under conditions shown in panels A & B above, in a separate set of preparations. In slices pre-exposed to two SDs (1^st and 2^nd SD), subsequent Zn/Pyr addition did not produce any increase in the amplitude of extracellular adenosine following SD (3^rd SD). No statistical difference was seen between the three SD challenges in the same slices (repeated measures ANOVA with Bonnferroni’s multiple comparison). In time-control slices not previously exposed to SDs (matching Fig. 7B), there was still a significant increase in extracellular adenosine accumulation by Zn/Pyr with SD (see also Fig. 3C). (*p<0.01 Student’s unpaired t-test, n=5 for each group).

Figure 7C shows extracellular adenosine measurements under these conditions of repetitive SD shown in Figure 7A&B. In slices pre-exposed to two SDs, subsequent Zn/Pyr addition did not produce any increase in the amplitude of extracellular adenosine following SD. This is consistent with the lack of effect of Zn/Pyr exposure in Figure 7A. In time-control slices not previously exposed to SDs (matching Fig. 7B), there was still a significant increase in extracellular adenosine accumulation by Zn/Pyr with SD. Exogenous adenosine application (25µM) following the end of the recording session produced similar decreases in fEPSPs under both SD pre-exposed and time-control conditions (final fEPSP amplitudes: 28.29±3.14 vs. 28.16±2.56% baseline, SD pre-exposed vs. time-control respectively, p=0.97, n=5 each). These results suggest that the loss of effectiveness of Zn/Pyr on fEPSP recovery is likely a consequence of decreased extracellular adenosine accumulation, rather than a change in sensitivity of synaptic transmission to adenosine, or other desensitization mechanisms.

DISCUSSION

General

This study assessed the contributions of [Zn²⁺]_i accumulation to the recovery from spreading depolarization (SD) in acute hippocampal brain slices. Transient [Zn²⁺]_i loading with pyrithione (Zn/Pyr) appeared enhance the metabolic burden following SD, measured by enhanced fEPSP suppression, prolonged extracellular DC shift durations, and increased extracellular adenosine accumulation. Transient [Zn²⁺]_i loading alone led to a slow run-down of fEPSPs, however when combined with SD, the cumulative metabolic challenge was significantly enhanced. Endogenous Zn²⁺ release and accumulation during SD propagation did not lead to a significant delay in fEPSP recovery. This could be explained by the need for sufficient time prior to SD for [Zn²⁺]_i accumulation to have significant additive effects, as well as competition with a rapid adaptive mechanism activated following SD that confers resistance to [Zn²⁺]_i effects during repetitive SDs in the brain slice model.

Additional adenosine accumulation following [Zn²⁺]_i loading together with SD

Previous work has demonstrated accumulation of extracellular adenosine and suppression of evoked fEPSPs under metabolically compromised conditions (Rudolphi et al. 1992, Masino et al. 1999). Such a mechanism appears to explain the additive effect of Zn²⁺ with SD in the current study, as extracellular adenosine accumulation was significantly enhanced by [Zn²⁺]_i loading, and the A1R antagonist DPCPX improved fEPSP recovery (Fig 3A). The adenosine measurements reported here are estimates of adenosine concentrations at the slice surface, determined by comparison with exogenous adenosine applications (Lindquist & Shuttleworth 2012). Concentrations of endogenous adenosine at synaptic sites within slices may be quite different, but regardless of absolute levels, the results here show significant functional increases in adenosine accumulation after SD in slices pre-loaded with Zn²⁺.

A significant run-down of fEPSPs was seen with [Zn²⁺]_i loading in the 60min time-control experiments where SD was not induced (Fig 1), and suggestions of this run-down could be seen even in the initial 10min washout period prior to SD challenges. It seems likely that A1R suppression also contributes to this slow run-down, and is additive with the metabolic challenge of SD. Curiously, a small transient increase in fEPSPs was seen upon the addition of Zn/Pyr (seen most easily in Fig 4). This increase was present during the addition of either pyrithione or histidine alone (data not shown, n=6 each), which suggests that extracellular or contaminating Zn²⁺ (see Discussion in Carter et al. 2011) may contribute by effects on either excitatory or inhibitory transmission (Xie et al. 1994, Huang et al. 2008, Kim et al. 2002).

Suppression of excitatory neurotransmission by adenosine acting on A1Rs has been widely considered neuroprotective, and following hypoxic challenges, fEPSP recovery was impaired in the presence of DPCPX or in mice lacking A1Rs (Johansson et al. 2001, Sebastiao et al. 2001). However in the current study of SD in normoxic conditions, we did not detect deleterious effects of DPCPX, and found instead that synaptic recovery was improved under all conditions tested. The majority of fEPSP suppression was mediated by A1R activation following 30µM Zn/Pyr. While fEPSP recovery was significantly improved by DPCPX with the 100µM Zn/Pyr challenge, some sustained fEPSP suppression was still seen (Fig. 3A). This suggests that under conditions of greatly increased metabolic inhibition, SD can result in either irrecoverable injury, out competition of DPCPX by excessive adenosine accumulation, or a combination of both. It remains to be determined whether adenosine accumulation following SD can contribute to delayed neuroprotection under other conditions, for example with clusters of SDs in ischemic conditions.

Mechanisms coupling [Zn²⁺]_i accumulation to adenosine accumulation

Several possibilities can be considered for mechanisms coupling [Zn²⁺]_i loading to adenosine accumulation. Sufficient time (many minutes) was required for [Zn²⁺]_i accumulation to have significant additive effects when combined with SD, consistent with a mechanism involving inhibition of metabolism or depletion of metabolic stores prior to SD. Based on the lack of effectiveness of selective inhibitors, PARP-1 or SIRT protein activation (see Introduction) did not appear to be significant contributors to the delayed fEPSP recovery following [Zn²⁺]_i loading together with SD (see Section 4). More prolonged Zn²⁺ exposures were utilized in studies which these pathways were identified in toxic Zn²⁺ effects (Sheline et al. 2000, Cai et al. 2006) and it is possible that postsynaptic [Zn²⁺]_i accumulation here was not sufficiently high enough, or of long enough duration to activate PARP-1 or SIRT pathways. An additional test of PARP-1 and SIRT pathway involvement could involve NAD⁺ supplementation (Sheline et al. 2000). However, this possibility was difficult to test experimentally in these slice experiments, as application of NAD⁺ alone was sufficient to suppress fEPSPs in a DPCPX-sensitive manner (data not shown). This effect may be due to its ability to mimic adenosine binding at A1Rs (Galarreta et al. 1993).

The lack of effect of pyruvate and lactate supplementation (see Section 4) suggest that any metabolic inhibition caused [Zn²⁺]_i loading is likely due to actions downstream of pyruvate utilization, within either the TCA cycle (Brown et al. 2000) or the mitochondrial electron transport chain (ETC) (Sharpley & Hirst 2006). Interestingly, fEPSP recovery was enhanced in the presence of the DMSO vehicle used for both the DPQ and sirtinol studies (compare Figs 4A&B). DMSO has been reported to have antioxidant properties (Sanmartin-Suarez et al. 2011), and it is possible that DMSO reduced reactive oxygen species (ROS) generation by Zn²⁺. Inhibition of ETC is established to lead to ROS production (Adam-Vizi 2005) and mitochondrial Zn²⁺ accumulation, impairment of mitochondrial function, and ROS production have all been reported following Zn²⁺ accumulation in neurons (Seo et al. 2001, Sensi et al. 1999, Dineley et al. 2005).

In addition to accumulation due to metabolic depletion, it is possible that other sources of adenosine accumulation could also contribute to effects of Zn²⁺ loading. Higher order metabolic precursors of adenosine, such as ATP release from astrocytes (Halassa & Haydon 2010) or cAMP release (Rosenberg & Dichter 1989) can be converted to adenosine extracellularly. It is not yet known whether Zn²⁺ may play a role in either of these, or other, mechanisms of extracellular adenosine accumulation.

Effects of [Zn²⁺]_i loading on the characteristics of SD

Longer DC shift durations during SD are associated with worse clinical outcomes in a variety of brain injuries that involve impaired metabolism (Hartings et al. 2011b, Dohmen et al. 2008, Lauritzen et al. 2011). Recent work in vitro has also shown that longer DC shifts are associated with longer fEPSP suppression in hippocampal slices (Lindquist & Shuttleworth 2012). We found here that [Zn²⁺]_i loading prior to SD significantly prolonged the duration of DC shifts (Fig 2), together with the effects on fEPSP recovery and adenosine accumulations discussed above. These results further the notion that longer DC shifts, both in vitro and in vivo, can be an indicator of larger metabolic challenges, and support a metabolic depletion mechanism for the Zn²⁺ effects described here.

Interestingly, there was no effect of [Zn²⁺]_i loading on SD propagation rate (Fig 2C), despite the fact that depletion of metabolic substrates can increase SD propagation rates (Seidel & Shuttleworth 2011). This suggests that either the degree of metabolic inhibition with [Zn²⁺]_i loading is mild prior to SD, or the nature of metabolic inhibition (i.e.. substrate removal vs. direct metabolic inhibition) determines effects on propagation rate.

Endogenous Zn²⁺ accumulation

We previously showed waves of synaptic Zn²⁺ release and intracellular accumulation following SD propagation (Carter et al. 2011). However, results from the present study suggest that synaptic Zn²⁺ release and subsequent postsynaptic accumulation does not lead to signs of progressive metabolic depletion, even when SDs were generated repetitively within the same slice (see Fig 6). One explanation for this lack of effect could be that endogenous [Zn²⁺]_i levels achieved following SD are much lower than those achieved during [Zn²⁺]_i loading with pyrithione. However, the results of Figure 7 also suggest the potential for a significant and rapid adaptation following SD in brain slices, making the tissue entirely resistant to a subsequent metabolic challenge with Zn²⁺ loading. SD is established to precondition tissues against subsequent ischemic challenges in vivo, although this effect has been tested only at relatively late time points (24hrs or greater) following SD induction (Yanamoto et al. 1998, Matsushima et al. 1996). It is not yet known whether the adaptation observed at early time points here shares a similar mechanism with ischemic preconditioning in vivo, or with PKC mechanisms linking intracellular Zn²⁺ increases to preconditioning in cultured cortical neurons (Aras et al. 2009).

The small increase in fEPSP recovery seen in the presence of TPEN, together with the lack of effect on fEPSP recovery in ZnT3 KO animals (see Fig 6), suggests that other endogenous, non-synaptic sources of Zn²⁺ can influence fEPSP recovery following SD. Potential contributors could include liberation of [Zn²⁺]_i from Zn²⁺-binding proteins (Aizenman et al. 2000) that may occur during conditions generated by SD, or from influx of contaminating Zn²⁺ present in the superfusate (see Discussion in Carter et al. 2011). However, the effect of TPEN on fEPSP recovery was quite small, and possibly counteracted by adaptive mechanisms following SD as described above.

Implications for injury

The results here emphasize potentially deleterious effects of intracellular Zn²⁺ accumulation, when combined with SD. Sustained [Zn²⁺]_i increases are known to occur in the days following stroke injury (Koh et al. 1996), and it seems likely that these sustained increases would make neurons more vulnerable to additional metabolic challenges of SDs that occur in the post-stroke brain (Selman et al. 2004, Hartings et al. 2003, Dreier 2011, Lauritzen et al. 2011). However, we have also shown that extracellular actions of Zn²⁺ can limit the propagation of SD (Aiba et al. 2012). Therefore, selectively targeting intracellular Zn²⁺ accumulation may be more effective in limiting neuronal injury and improving outcome following several brain injury models, particularly those that involve SD. However it is important to also emphasize that clusters of repetitive SD have been associated with poor clinical outcomes under some conditions (Dreier 2011, Lauritzen et al. 2011, Hartings et al. 2011a) and a rapid adaptation following SD appears to protect tissues against the additive metabolic burden of intracellular Zn²⁺ (Figure 7). Therefore deleterious metabolic effects of intracellular Zn²⁺ accumulation may not be significant during repetitive SDs clusters, and under such conditions, extracellular effects of Zn²⁺ may dominate and limit the spread of these events (Aiba et al. 2012). These and other divergent actions of intracellular and extracellular Zn²⁺ are likely contributors to different effects of Zn²⁺ chelation that have been reported in experimental injury settings (Kitamura et al. 2006, Lee et al. 2002).

Abbreviations used

A1R

adenosine A1 receptor

ACSF

artificial cerebral spinal fluid

Ado

adenosine

CA1

Cornu Ammonis 1

DC

direct current

DMSO

dimethyl sulfoxide

DPCPX

8-cyclopentyl-1,3-dipropylxanthine

DPQ

3,4-dihydro-5-[4-(1-piperidinyl)butoxyl]-1(2H)-isoquinolinone

fEPSP

field excitatory post-synaptic potential

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

IOS

intrinsic optical signal

KGDHC

α-ketoglutarate dehydrogenase complex

KO

knock out

PARP-1

poly(ADP-ribose) polymerase 1

Pyr

pyrithione

SD

spreading depolarization

TCA cycle

tricarboxylic acid cycle

TPEN

N,N,N’,N’-tetrakis(2-pyridylmethyl)-ethylenediamine

[Zn²⁺]_i

intracellular Zn²⁺

Zn/Pyr

ZnCl₂ together with pyrithione