- Email: [email protected]
Evidence for constitutive protein synthesis in hippocampal LTP stabilization
Neuroscience 246 (2013) 301–311
EVIDENCE FOR CONSTITUTIVE PROTEIN SYNTHESIS IN HIPPOCAMPAL LTP STABILIZATION A.-K. ABBAS *
inhibitors (PSIs) during a critical time window (often ±30 min) around the LTP induction event makes an LTP, that would otherwise be long-lasting, to vanish within few hours, i.e., prevents late-phase LTP. Thus, an LTP-inducing tetanization can result in late-phase LTP only if this tetanization triggers synthesis of proteins stabilizing LTP. However, results from other studies have questioned this idea. For example, this effect of PSIs was not observed when also protein degradation was blocked indicating that constitutively expressed proteins may suffice for LTP stabilization (Fonseca et al., 2006). We and others have also recently failed to observe any effect of protein synthesis inhibition on LTP during its first 7–8 h (Abbas et al., 2009; Villers et al., 2012). Moreover, the stability of LTP following a given induction protocol may depend critically on experimental conditions, such as slice preparation and recovery conditions (for references, see Villers et al., 2012). That the state of the slice may affect LTP raises the question to what extent procedures affecting the protein content and/or the state of the proteins in the slice may affect LTP stabilization. Thus, can protein depletion per se result in an impaired ability of the synapses to undergo the structural/functional rearrangements necessary for stable LTP? In the present study we have introduced a long (4 h) pre-application of PSI to examine whether this procedure may result in less stable LTP. In another set of experiments we applied a mixture of anisomycin and cycloheximide (CHX) for 30 min before LTP induction throughout the time-course of recorded LTP. Furthermore, we also transiently applied hydrogen peroxide (H2O2) to the slice in order to induce rapid protein modification/degradation. Our findings suggest that proteins synthesized in response to the LTP-inducing stimulation are unlikely to be critical for the subsequent stabilization of LTP.
Institute of Neuroscience and Physiology, University of Gothenburg, P.O. Box 433, SE-40530 Gothenburg, Sweden
Abstract—The notion that blockade of constitutive protein synthesis underlies the effect of protein synthesis inhibitors (PSIs) on long-term potentiation (LTP) stabilization was examined using the rat hippocampal CA3–CA1 synapse. Using a biochemical assay we found protein synthesis rate largely recovered 1 h after wash-out of cycloheximide (CHX). Nonetheless, a 4-h CHX application followed by wash-out 1 h prior to LTP resulted in a significant decrement of LTP stabilization. Wash-out initiated just prior to LTP, thus extending protein synthesis inhibition well into the post-LTP period, resulted in no further effect on LTP. However, short pre- and continuous post-tetanization application of PSIs failed to influence LTP persistence for up to 7 h. Addition of hydrogen peroxide (H2O2) 5–25 min following LTP induction resulted in parallel depression of potentiated and non-potentiated inputs, leaving LTP seemingly unaltered. However, in the presence of cyxloheximide the H2O2 application resulted in a significant reduction of LTP. In conclusion: LTP stabilization was impaired by pre-LTP application of protein synthesis inhibition but not by postLTP application unless the slices were exposed to oxidative stress. We submit that these results favor the notion that constitutive rather than triggered protein synthesis is important for LTP stabilization. Ó 2013 IBRO. Published by Elsevier Ltd. All rights reserved.
Key words: anisomycin, cycloheximide, neuronal plasticity, oxidative stress, protein synthesis, synapse.
INTRODUCTION Long-term potentiation (LTP) is a cellular model for some forms of learning and memory (Martin et al., 2000; Lynch, 2004). In common with memory LTP has been differentiated temporally into an early vs. late phase, the latter believed to require the synthesis of novel proteins for its expression. In fact, a large number of studies have shown that application of protein synthesis
EXPERIMENTAL PROCEDURES Animals The experimental biomedicine (EBM) animal facility is fully accredited by the Swedish Council for Laboratory Animals. Fourteen- to 22-day-old Albino rats (strain Sprague–Dawley) of mixed sexes were obtained from Charles River (Scanbur AB, Sollentuna, Sweden) and were prior to the experiment kept group-housed. By using a multi-recording chamber system allowing four slices to be studied in parallel, the number of animals used was kept at a minimum.
*Tel: +46-768-235900; fax: +46-31-7863840. E-mail address: [email protected] Abbreviations: ACSF, artificial cerebrospinal fluid; AP-5, D-( )-2amino-5-phosphonopentanoic acid; CHX, cycloheximide; DMSO, dimethysulfoxide; DPCPX, 8-cyclopentyl-1,3-dipropylxanthine; fEPSP, field excitatory postsynaptic potentials; FeSO4, ferrous sulfate; H2O2, hydrogen peroxide; LTP, long-term potentiation; PSIs, protein synthesis inhibitors; TBS, theta-burst stimulation; TCA, trichloroacetic acid.
0306-4522/13 $36.00 Ó 2013 IBRO. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.neuroscience.2013.05.011 301
302
A.-K. Abbas / Neuroscience 246 (2013) 301–311
In vitro electrophysiology and drugs preparation The animals were killed in accordance with the guidelines of the local ethics committee of University of Gothenburg. They were deeply anesthetized with isoflurane (Baxter Medical AB, Oslo), decapitated and the brain was removed. Transverse hippocampal slices (400-lm-thick) were cut using a McIlwain-type tissue chopper and submerged in a holding chamber containing roomtempered artificial cerebrospinal fluid (ACSF) composed of (in mM): 119 NaCl, 2.5 KCl, 2.0 CaCl2, 2.0 MgCl2, 26 NaHCO3, 1.0 NaH2PO4, and 10 D-glucose, equilibrated with 95% O2 and 5% CO2. After pre-incubation for at least 90 min, slices were transferred as needed to one or several recording chambers where they were kept submerged and covered with nylon net to prevent their movement. The recording chamber consisted of a circular well of low volume (1.5–2 ml), and the solution was recycled (1.5–2.0 ml/min) using a peristaltic pump (Ismatec, Labinett Lab AB, Sweden). The perfusion solution was the same as in the holding chamber except for 2.5 mM CaCl2 and 1.3 mM MgCl2. Experiments were performed at 31 °C. Extracellular field potentials were recorded with a glass micropipette filled with 1 M NaCl (resistance 3–5 MX) positioned in the middle of the CA1 stratum radiatum. Microelectrodes were pulled from microfiber (o.d. 1.5 mm, i.d. 0.86 mm, Warner Instruments, LLC, Hamden, CT, USA) capillary tubing. To stimulate independent inputs to the same cell population, two monopolar tungsten stimulating electrodes (0.1 MX; World Precision Instrument, Inc., Sarasota, FL, USA) were positioned on either side of the recording microelectrode. Their positions were arranged so that the same amount of stimulating current evoked field potentials of similar magnitude. Stimuli were delivered to the commissural–Schaffer collateral afferents as 100-ls negative constant-current pulses (20–50 lA) using a programmable pulse generator. While one synaptic input (test pathway) was employed to induce and monitor LTP, the other one (control pathway) was used to verify the stability of basal synaptic transmission. Pulses were delivered alternately to the two pathways, usually at a rate of once every 40 or 60 s, with stimuli to the two pathways separated by 20 or 30 s, respectively. Baseline field excitatory postsynaptic potentials (fEPSPs) were recorded for 60–90 min to ensure stability of responses. LTP was induced generally by theta-burst stimulation (TBS) composed of three trains (inter-train interval 5 s), each consisting of 10 bursts of four pulses at 100 Hz, with an inter-burst interval of 200 ms (Larson et al., 1986). In one group of experiments, LTP was induced by three tetanus trains (100 Hz, 1 s, inter-train interval 5 s). The PSIs anisomycin (2-[p-methoxybenzyl]-3,4, pyrrolidinediol-3-acetate) and CHX 4-{(2R)-2[(1S,3S,5S)-3,5-dimethyl-2-ococyclohexyl]-2-hydroxyethyl} piperidine-2,6-dione, H2O2 (3%), ferrous sulfate (FeSO4) and dimethysulfoxide (DMSO) were all purchased from Sigma–Aldrich (St. Louis, MO, USA). 8-Cyclopentyl-1,3dipropylxanthine (DPCPX) and D-( )-2-amino-5phosphonopentanoic acid (AP5) were obtained from
Tocris Bioscience (UK) or Ascent Scientific Ltd (UK). [3H]leucine was obtained from Amersham, Buckinghamshire, UK. Milli-Q deionized water (Millipore, Bedford, MA, USA) was used in all preparations of buffers and solutions. Other chemicals used were all of highest grade commercially available. For biochemical assay, anisomycin, or CHX, was added to ACSF to different final concentrations derived from different stocks prepared in distilled water at the time of each experiment. For electrophysiological experiments, CHX and DPCPX were first dissolved in DMSO to a concentration of 100 and 3 mM, respectively, and stored at 18 °C until diluted in physiologic buffer to final concentration of 100 lM and 50 nM, respectively. The largest final concentration of DMSO was 0.1%, which had no effect on basal synaptic transmission as verified by lack of effect on the baseline fEPSP when added as vehicle. Stocks of AP5, anisomycin and FeSO4 were prepared in distilled water and diluted at final concentrations of 50 lM, 25 mM and 100 lM, respectively. The rational for FeSO4 addition was to enhance the generation of hydroxyl radical from H2O2 via a Fenton-like reaction. Biochemical assays Effect of PSIs on protein synthesis was measured by [3H]leucine incorporation (Lipton and Heimbach, 1978) in whole slices. Hippocampal slices from 13- to 24-dayold rats were maintained under similar conditions as in the electrophysiological work but without stimulation. Slices were put on a multiwall plastic dish (Corning Incorporated, Corning, NY, USA) and assigned to a PSI group or a control group in an interleaved manner to minimize inter-slice variability with respect to weight and metabolism. To test the effect of PSIs on rate of protein synthesis in dose-dependent manner, two slices per group were pre-incubated for 10 min with the drug (or vehicle, if necessary) before addition of [3H]leucine; final activity 0.5–1 lCi/ml. Uptake and incorporation of leucine into proteins were allowed to proceed for 50 min either with or without PSI. To test PSI reversibility, a corresponding total duration of pre-incubation with the drug (60 min) was followed by different pre-incubation intervals (30, 60 or 90 min) in drug-free solution before allowing slices to be incubated in a [3H]leucinecontaining solution for additional 50 min. Radioactive amino acid incorporation was then terminated by the addition of ice-cold saline, and the slices were subsequently lysed in 1 ml of 5 mM NaOH. After protein purification, incorporation of leucine into trichloroacetic acid (TCA)-precipitable proteins was measured in a scintillation counter (LKB Wallace, 1219 Rackbeta, Finland). Percentage inhibition of leucine incorporation produced by drug treatment was calculated by comparing counts in treated slices with those of control ones. To investigate whether protein content was influenced by the addition of H2O2, two groups of slices were incubated with tritiated leucine for 50 min at 31 °C following a pre-incubation period equivalent to that for the electrophysiological experiments. The slices were subsequently washed by continuous perfusion with
A.-K. Abbas / Neuroscience 246 (2013) 301–311
303
ACSF for 30 min before H2O2/FeSO4 was applied to one group of slices for 20 min at concentrations identical to those used for the electrophysiological experiments. Both control and H2O2/FeSO4-exposed slices were subsequently rinsed and frozen at different time intervals following the exposure. Percentage of leucine incorporation produced by H2O2/FeSO4 was calculated by comparing counts in treated slices with those of control slices. Data analysis Electrophysiological data were collected using in-house developed electronic equipment based on an Eagle Instruments multifunction board, capturing signals from up to four parallel experiments. Signals were amplified, filtered, digitized and transferred to a digital computer for on-line and off-line analysis. Off-line data analysis was performed using a combination of Quickbasic (Microsoft) and Igor (Wavemetrics, Inc.) programs. fEPSPs were measured using an early time window of 1–2-ms duration positioned just after the presynaptic volley. The area under the curve was used as a measure of fEPSP size. Alternatively, the fEPSP can be quantified by a slope measurement, shown to give similar results (Dozmorov et al., 2003). The area measurement was generally preferred in the present study due to less noise and better immunity to volleyEPSP overlap. Slices in which rundown of the control fEPSP was by more than 30% at the end of experiments (10 h of recording) were excluded (Volianskis and Jensen, 2003). For time course plots, LTP values were averaged over 7 min (7–10 successive responses depending on test frequency). For each successive measurement, the mean, standard deviation, and standard errors were calculated for all the experiments. fEPSP (mV/ms) was calculated and data for each experiment were normalized relative to baseline recording. As normality assessment (Q–Q plot and Shapiro–Will test) revealed non-Gaussian distribution of groups, nonparametric Mann–Whitney U test was conducted (SPSS for windows, Version 18.0) when data were compared between groups (drug vs. control). Alternatively, Student’s t-test was used to assess comparisons when the distribution of groups was Gaussian. Numbers of experiments are indicated by n. Values of p < 0.05 were considered to represent significant differences.
RESULTS Effect of prolonged post-induction application of PSIs on LTP We reported previously (Abbas et al., 2009) that anisomycin or emetine applied from 30 min prior to 30 min after the LTP induction event, thus covering the proposed critical interval of triggered protein synthesis (Frey and Morris, 1997, 1998), failed to affect LTP recorded for up to 8 h. As shown in Fig. 1, we also observed such failure when wash-out was omitted, the tetanization resulting in seemingly equally stable LTP in
Fig. 1. Lack of effect on LTP by continuous application of PSIs. LTP was induced by three trains of tetanus (100 Hz, 1- and 5-s inter-train interval) in one of two separate pathways, using the other pathway as control. (A) LTP induced in non-treated slices. (B) LTP induced in slices in the presence of anisomycin (25 lM) and cycloheximide (70 lM) as indicated by the horizontal bar. (C) LTP shown in A and B are plotted superimposed after drift-compensation (test/control ratio) indicating no effect of the PSI application. Insets in A and B show superimposed fEPSPs of control and test pathways taken before, 4 h after, and 7 h after induction of LTP. Arrows indicate time of LTP induction. Each value in the graph represents the average fEPSP (±S.E.M.) sampled over 10-min periods.
the treated group of slices (Fig. 1B, n = 5) as in the non-treated group (Fig. 1A, n = 7). It can be noted that the drift of the control pathways was small and did not differ between the two groups suggesting that the prolonged protein synthesis inhibition by itself did not affect the synaptic transmission. To directly compare the LTP we compensated for the drift by computing the ratio of test to control responses. As shown in Fig. 1C, where
304
A.-K. Abbas / Neuroscience 246 (2013) 301–311
these drift-compensated LTP are shown superimposed, they were closely overlapping. Thus, in contrast to Ris et al. (2009); but see Villers et al. (2012), we failed to observe that a prolongation of the post-induction drug application results in a more efficient blockade of LTP. In the above experiments, two PSIs were used together (anisomycin, 20 lM and CHX, 80 lM), avoiding emetine which has nonspecific effects when applied for a long duration (Abbas et al., 2011). Nonetheless, in view of our negative results we realized the need to control for the efficacy of these drugs under conditions that faithfully match those of our present electrophysiological studies. Moreover, since in the remaining part of our study the protein synthesis blockade should occur only prior to the LTP induction event, we needed to establish the time course of reversal from the protein synthesis blockade following drug wash-out. Determination of potency of the PSIs Fig. 2A shows percentage inhibition of protein labeling using 50-min incubation with [3H]leucine and a variable concentration of anisomycin, indicating an IC50 value in the order of 50 nM. Even the lowest concentration, 10 nM, resulted in a measurable inhibition (about 20%) and a saturating inhibition (>90%) was achieved with concentrations P 5 lM. Doses in 10–25 lM range are commonly used to inhibit protein synthesis and were previously employed by us (Abbas et al., 2009, 2011). As evidenced by the results in Fig. 2A, such concentrations of anisomycin should be capable of efficient inhibition of protein synthesis. Fig. 2B shows the inhibition profile obtained for CHX, indicating an IC50 value of 300 nM, substantially higher than the corresponding value for anisomycin. Saturation (95% inhibition) was reached by concentrations of P40 lM. To ascertain that the mixture of CHX and anisomycin in the above experiments resulted in an optimal protein synthesis inhibitory effect, we conducted a [3H]leucine incorporation assay. As shown in Fig. 2C, the mixture of anisomycin and CHX resulted in 95 ± 0.4% (n = 4) protein synthesis inhibition. Notably, the level of protein synthesis inhibition produced by either anisomycin (Fig. 2A) or CHX (Fig. 2B) alone was not significantly different from that produced by the mixture (anisomycin, 93 ± 1%, n = 8; CHX, 94 ± 0.6%, n = 8). Reversibility of protein synthesis blockade following drug wash-out [3H]leucine assays were used to test the reversibility of anisomycin- and CHX-induced protein synthesis blockade. The inhibitor was first applied for 60 min, this application followed by wash-out and variable intervals (0–90 min) before [3H]leucine (without PSI) was applied for a 50-min period (Fig. 3). The experiment was finalized by analyzing protein content (see Experimental procedures, and the illustrated scheme in Fig. 3A). The histogram data in Fig. 3B shows that both anisomycin and CHX at 0-min wash-out essentially fully blocked
Fig. 2. Dose-dependent effect of PSIs on [3H]leucine incorporation into hippocampal slices. (A) Slices, at each concentration of anisomycin used, were first pre-incubated either with normal solution (control) or with anisomycin (test) for 10 min before adding radioactive leucine to these solutions for further 50 min. (B) Same as in A but using cycloheximide. The n values (equal for test and control at a given concentration) ranged between four and eight for anisomycin experiments, and between four and 12 for the cycloheximide experiments (part of these data has been published in Abbas et al., 2011). (C) Maximal level of protein synthesis inhibition (±S.E.M.) obtained by cycloheximide (CHX; n = 8), anisomycin (Ani; n = 7), and a mixture of Ani plus CHX (n = 4), indicating no significant difference between these treatments.
protein synthesis (as was the case with overlapping PSI and [3H]leucine application shown above). When longer wash-out periods (30, 60, and 90 min) were allowed prior to [3H]leucine application the CHX-induced inhibition decreased substantially. For instance, after 30 and 60 min of wash-out, protein synthesis rate was up to 68% and 80%, respectively, of control. The protein synthesis inhibition induced by anisomycin was less reversible, the protein synthesis rate no more than 10% of control even after 90 min of wash-out.
Effect of prolonged pre-TBS application of CHX, with pre-TBS washout, on LTP In view of the poor reversibility of the anisomycin-induced blockade, only CHX was used to test the effect of an extended period of pre-TBS protein synthesis inhibition on LTP. In the design of these experiments the period of drug application ought to be long enough to cause a significant depletion of a wide range of constitutively
A.-K. Abbas / Neuroscience 246 (2013) 301–311
305
normal and drug groups, respectively; U = 5; r = 0.6; p < 0.05; Mann–Whitney U test). Effect of prolonged pre-TBS application of CHX without pre-TBS washout on LTP
Fig. 3. Time-course of PSI reversibility. (A) Schematic illustration of the experiments. Anisomycin or cycloheximide was first applied for 60 min, followed by a variable washout interval before [3H]leucine assays was conducted for further 50 min. The experiment was eventually finalized by analyzing protein content. Four time points were estimated with the one with 0 wash-out was used as control (a). Values from three washout intervals: 30 min (b), 60 min (c) and 90 min (d) were compared with the control group (a). (B) Assessing time-course of reversibility of anisomycin (25 lM; n = 4 for each time point; black bars) or cycloheximide (40 lM; n = 4 for each time point; gray bars) using tritiated leucine incorporation showed that while washout interval following cycloheximide treatment requires >30 min for protein synthesis recovery to reach 80%, anisomycin washout was not followed by protein synthesis recovery up to 90 min. The group mean percentage (±S.E.M) for each time point of washout following protein synthesis inhibition computed for n slices in the protein synthesis inhibitor group and n slices in the control group without washout interval (dashed line; 100%).
synthesized proteins. Given that proteins involved in synaptic plasticity may have half-lives up to 2–3 h in vitro (Sajikumar and Frey, 2004), that mammalian extracellular proteins have half-lives in the order of 4 h (Hesse et al., 1984), and that tag lifetime has been reported to be 1–2 h (Frey and Morris, 1997, 1998; but see Fonseca et al., 2004), we chose a 4-h CHX application paradigm, followed by 1 h of wash-out before LTP induction and subsequent 7 h of LTP recording. This paradigm was experimentally feasible since total recording time in our hands can generally extend up to 12–13 h. These results obtained using CHX was compared with those obtained using non-treated slices in which the LTP-inducing TBS trains were given following 5 h of baseline stimulation. Fig. 4 shows LTP obtained in non-treated (A) and treated (B) slices, respectively, without compensation for control input drift, and the corresponding drift-compensated LTP are shown superimposed in (C). It can be noted that while LTP values of non-treated and treated slices did not differ when measured 1 h after induction (95 ± 9% vs. 91 ± 15%; U = 16; r = 0.09; p > 0.05; Mann–Whitney U test) they did significantly differ 5 h later (47 ± 13% vs. 10 ± 9%; U = 4; r = 0.65; p < 0.05; Mann–Whitney U test). The drift-compensated LTP revealed a significant difference at 8 h (60 ± 5% vs. 40 ± 4%, for
In initial experiments prior to the realization of the slow reversibility of the protein synthesis inhibition, LTP was induced soon after the onset of CHX wash-out. Thus, in these experiments protein synthesis was not only blocked for 4 h preceding LTP, but to a substantial extent also for roughly 1 h post-TBS, thus covering much of the critical period believed to exist for the triggered protein synthesis. Fig. 5 shows that also under this condition the LTP of non-treated (A) and treated (B) slices did not differ when evaluated 1 h after induction (110 ± 8% (n = 8) vs. 114 ± 11% (n = 14); U = 42; r = 0.2; p > 0.05; Mann–Whitney U test) but differed significantly 5 h later (63 ± 10% vs. 35 ± 12%; U = 22; r = 0.5; p < 0.05; Mann–Whitney U test). Values obtained as test/control input ratio (Fig. 5C) at 6 h were in order of 190 ± 7% vs. 161 ± 6% (U = 15; r = 0.6; p < 0.005; Mann–Whitney U test). The parallel findings evident both from the original data in Figs. 5A, B and the drift-compensated ones in Fig. 5C were a consequence of quantitatively similar rundown of the control pathway among the two groups. Effect of H2O2 application on synaptic transmission and on LTP To induce protein modification/degradation following LTP induction H2O2 (5 mM) together with FeSO4 (100 lM) was applied for 20 min, starting 5 min following the LTP induction event. To reduce some unwanted side effects, the experiments were performed in the presence of the adenosine A1 receptor antagonist DPCPX (50 nM), and the NMDA receptor antagonist AP5 (50 lM) was added after the LTP induction. Fig. 6A shows that H2O2/FeSO4 application had a substantial immediate depressing effect on the control inputs (n = 6) that remained to much the same extent after H2O2/FeSO4 wash-out and throughout the 7 h post-LTP induction recording period. This depressive effect on the control inputs was seemingly mirrored by a similar effect on the LTP of the potentiated inputs (in order of 141 ± 13% and 126 ± 14% at 2 and 7 h, respectively; Fig. 6B). In Fig. 6C LTP exhibited as ‘‘drift-compensated’’ is shown to be stable indicating a parallel depression to that of the control input. Measurements of the presynaptic fiber volley indicated no effects of H2O2/FeSO4 application on axon excitability (data not shown). CHX reduces LTP H2O2/FeSO4 was subsequently applied to slices in the presence of CHX (80–100 lM), whose application started 30 min before LTP induction and kept throughout the 7-h recording period. As shown in Fig. 7A, B both control and potentiated inputs (n = 11) became depressed by the H2O2/FeSO4 application in much the same manner as in the absence of CHX. However, in
306
A.-K. Abbas / Neuroscience 246 (2013) 301–311
Fig. 4. Effect of long pre-incubation with cycloheximide prior to LTP induction. (A) LTP was induced by a TBS paradigm (three trains of 10 bursts of four pulses at 100 Hz, repeated with a burst frequency 5 Hz, 5-s inter-train interval) in one of two separate pathways, using the other as control (n = 6). TBS trains were given following 5 h of baseline stimulation. (B) Same as in (A), but in slices that had been pre-incubated with cycloheximide (100 lM) for 4 h, the cycloheximide wash-out beginning 1 h prior to LTP induction (n = 6). (C) LTP shown in (A, B) is plotted superimposed after drift-compensation (test/control ratio) indicating less stability of LTP in PSI-treated slices. Insets in A and B show superimposed fEPSPs of control and test pathways taken before, 4 h after, and 7 h after induction of LTP. Arrows indicate time of LTP induction. Each value in the graph represents the average initial area of fEPSPs (±S.E.M.) sampled over 8 min periods.
the presence of CHX LTP appears disproportionally affected (in order of 121 ± 12% and 90 ± 8% at 2 and 7 h, respectively; Fig. 7B). We thus computed the ‘‘driftcompensated’’ LTP by taking the ratio test/control responses to assess the effect of CHX on LTP itself. In Fig. 7C these ‘‘drift-compensated’’ LTP are shown superimposed, indicating little effect of protein synthesis inhibition during the first 3–4 h of LTP. However, thereafter, LTP in CHX-treated slices decays substantially implying a role for protein breakdown in LTP stabilization following the transient H2O2/FeSO4
application. Significant LTP values were observed at 7 h following LTP induction amounting 195 ± 16% vs. 145 ± 8% (t = 3.2078; df = 15; p < 0.05; Student’s t test) for H2O2/FeSO4 and H2O2/FeSO4/CHX group, respectively. H2O2 causes a significant reduction in the newly synthesized proteins We also examined the effect of H2O2 (5 mM, 20 min) on the turnover of newly synthesized proteins. The
A.-K. Abbas / Neuroscience 246 (2013) 301–311
307
Fig. 5. Effect of long pre-incubation with cycloheximide prior to LTP induction. (A) LTP was induced by a TBS paradigm (ten 4-impulse 100 Hz trains at 5 Hz, repeated thrice at 0.2 Hz) in one of two separate pathways, using the other as control (n = 6). TBS trains were given following 4 h of baseline stimulation. (B) Same as in (A), but in slices that had been pre-incubated with cycloheximide (100 lM) for 4 h, the cycloheximide wash-out beginning immediately prior to LTP induction. (C) LTP shown in (A, B) is plotted superimposed after drift-compensation (test/control ratio) indicating less stability of LTP in PSI-treated slices. Insets in (A, B) show superimposed fEPSPs of control and test pathways taken before, 4 h after, and 7 h after induction of LTP. Arrows indicate time of LTP induction. Each value in the graph represents the average initial area of fEPSPs (±S.E.M.) sampled over 8 min periods.
histogram data in Fig. 8 show that the percentage of newly synthesized radiolabeled proteins 1 h following H2O2 washout was nearly stable compared to control values (100%). However, at 3 h following washout it decreased to moderate but significant degree amounting 85 ± 4% (p < 0.05; cf. Grune et al., 1995, 1996).
DISCUSSION Three main results emerge from the present study. First, that an extended pre-TBS period of protein
synthesis blockade destabilizes LTP maintenance simulating a partial blockade of ‘‘L-LTP’’. Second, continuous protein synthesis application after LTP induction did not cause an observable effect on LTP stabilization. Third, that after a transient post-TBS application of H2O2, but not in control conditions, protein synthesis inhibition covering the ‘‘critical’’ time window (first hour post-TBS) could destabilize LTP. As will be detailed below, we believe these results favor the idea that the supply of constitutively expressed, rather than of tetanization-triggered,
308
A.-K. Abbas / Neuroscience 246 (2013) 301–311
Fig. 6. Effect of oxidative stress on basal transmission and on LTP. Slices were in the presence of DPCPX and AP5 exposed to a 20 min application of hydrogen peroxide (H2O2) plus FeSO4. The rational for adding FeSO4 was to enhance hydroxyl radical formation via a Fenton-like reaction. (A) Control pathway. Note the immediate and lasting depression induced by this application. (B) Test pathway activated by LTP-inducing TBS (ten 4-impulse 100 Hz trains at 5 Hz, repeated thrice at 0.2 Hz) 5 min prior to the hydrogen peroxide (H2O2) plus FeSO4 application. Each value in the graph represents the average initial area of fEPSPs (±S.E.M.) sampled over 5-min periods. Insets in (A, B) show superimposed field EPSPs taken at times indicated in the graphs. (C) The ‘‘drift-compensated’’ LTP (test/ control ratio). Note that the magnitude and stability of LTP are as much as normally observed (cf. Fig. 5C) implying that the effect of hydrogen peroxide on either input is due to a reduction in the basal synaptic transmission.
proteins determines the stability of LTP maintenance at least within its first 8 h. Potency and time-course of protein synthesis inhibition For the interpretation of the present results it was vital that the pre-TBS protein synthesis blockade was largely restricted to the pre-TBS time period. At the
Fig. 7. Effect of cycloheximide (CHX) on basal transmission and on LTP in slices exposed to oxidative stress. The results shown in Fig. 6 are re-plotted in this figure (open circles) together with results obtained in slices had been exposed to cycloheximide (100 lM) throughout the experiment (starting 30 min prior to LTP induction) (closed circles). This set of experiments is also performed in the presence of DPCPX and AP5 at similar concentrations and time points. Note that while the field EPSPs of the control pathway (A) are little affected by the presence of cycloheximide, the field EPSPs of the tetanized pathway (B) are more affected, especially in the later part of the LTP. Insets in (A, B) illustrate superimposed fEPSP taken at times indicated in the graphs. In (C), where the drift-compensated LTP observed in the presence of cycloheximide is shown superimposed with that obtained in the absence of this drug, shows that cycloheximide impairs LTP stability in slices exposed to oxidative stress.
concentrations presently used, anisomycin and CHX were found to inhibit protein synthesis by 90–95%, but how reversible was this inhibition following drug washout? Here we found, using a [3H]leucine assay, that while both drugs had fast onset times (a few minutes), in the case of anisomycin no reduction of the inhibitory
A.-K. Abbas / Neuroscience 246 (2013) 301–311
309
Constitutive proteins in LTP stabilization
Fig. 8. Time course of protein labels percentage in hippocampal slice following [3H]leucine and H2O2/FeSO4. Pooled values (±S.E.M.) from eight and four experiments, for 1- and 3-h time points, respectively, compared to their own normalized control slices (broken line). The data reveal a significant decrease in the labeled proteins at 3 h but not 1 h following peroxide treatment. Each experiment is average of two slices.
effect was observed even after nearly 2 h of washout. This result agrees with that of Villers et al. (2012) who showed that anisomycin even after 4 h of wash-out inhibited protein synthesis rate by more than 80%. On the other hand, the CHX effect was reversible but still required about 1 h of wash-out for the synthesis rate to rise to 80% of control. In our pre-TBS experiments CHX was used with wash-out starting 1 h prior to the LTP induction, implying that some degree of protein synthesis inhibition would still extend into the ‘‘critical’’ time window for de novo synthesis of proteins believed to stabilize LTP. However, since in separate experiments in which the wash-out was started just before LTP induction instead of 1 h before did not result in a less stable LTP, it seems safe to conclude that the effect of our pre-TBS application is related to a pre-TBS rather than to a post-TBS protein synthesis blockade.
No role for triggered protein synthesis In our previous work (Abbas et al., 2009), application of neither emetine nor anisomycin for ±30 min around the time of LTP induction was found to influence LTP maintenance. In consideration of the possibility that the period of triggered de novo protein synthesis critical for LTP maintenance might vary among different studies depending on experimental conditions, in initial experiments we extended the post-tetanization protein synthesis blockade by omitting drug wash-out. We also applied a cocktail of anisomycin and CHX to maximize the inhibitory effect on protein synthesis. Despite these efforts, which considering our later observed high level of inhibition by each of these drugs alone and poor reversibility of the anisomycin-induced inhibition, may be found somewhat exaggerated, no de-stabilization of LTP maintenance was observed. Thus, under our experimental conditions triggered protein synthesis is not critical for the stabilization of LTP at least within its first 8 h. This lack of effect of prolonged posttetanization protein synthesis blockade was recently also reported by Villers et al. (2012) using older rats (6– 10 weeks), indicating that our use of young rats (2– 3 weeks) has not been decisive for obtaining this result.
The idea that LTP is stabilized beyond its first few hours by proteins whose synthesis is triggered by the LTP-inducing high-frequency stimulation rests almost entirely on the observation from numerous studies that application of PSIs during the ‘‘critical’’ time window results in an LTP that is largely decayed within few hours. The nature of these proteins and in what manner they should stabilize LTP are questions essentially unanswered (see e.g. Discussion in Villers et al., 2012). In general terms these proteins should ‘‘cement the synaptic change initiated during early phase LTP and contribute to activitydependent structural changes’’ (Fifkova et al., 1982; described by Abraham and Williams, 2008). In fact, the findings reported by Fifkova et al. (1982) were problematic in term that early (4 min) but not late (90 min) entorhinal spine morphological changes were prevented by administration of anisomycin in vivo. However, neither an early phase LTP expression based on e.g. CaMKII autophosphorylation nor activity-dependent structural changes will necessarily need new proteins. For example, rapid morphological changes such as persistent spine enlargement were observed within less than 1 min following glutamate receptors stimulation (Matsuzaki et al., 2004). Furthermore, emergence of new spines was documented to occur 30–60 min following high-frequency stimulation (Maletic-Savatic et al., 1999; Toni et al., 1999) or LTP induction (Engert and Bonhoeffer, 1999). Although there are no reports, up to our knowledge, confirming whether late morphological changes associated with LTP in vitro might be achieved without the requirement for triggered de novo protein synthesis, dendritic and/or neighboring macromolecules redistribution might not be ruled out. In this context, Tsuriel et al. (2006) showed that in cultured rat hippocampal neurons the dynamics of molecules, such as synapsin and ProSAP2, at individual CNS synapses may be dominated primarily by the continuous exchange and redistribution among nearby synapses lasting for up to 6 h, whereas protein synthesis and degradation may constitute slower, second-order processes that serve to maintain and regulate the size of local, shared pools of synaptic matrix proteins. Additional data confirmed that these dynamics are not unique to synapsin I and ProSAP2 but are characteristic of synaptic cytomatrix protein in general (for references, see Tsuriel et al., 2006). Some studies have also previously shown that LTP, albeit under specific experimental circumstances, can be stable without a tetanization-triggered synthesis of new proteins. Thus, if test pulse stimulation is interrupted post-tetanus or if protein degradation is simultaneously inhibited, LTP is maintained. Such results would then rather point to the necessity of a sufficient supply of proteins synthesized prior to LTP induction for stabilization of LTP, at least within its first 8 h. Our partial de-stabilization of LTP by a prolonged pre-TBS protein synthesis blockade would then concur with such an interpretation of the protein synthesis dependence of LTP. Villers et al. (2012) did, in contrast to us, not find any effect of a 4-h pre-application of anisomycin on LTP recorded for up to 8 h. It should
310
A.-K. Abbas / Neuroscience 246 (2013) 301–311
then first be noted that we observed but a partial de-stabilization of LTP, which implies that any difference in experimental conditions such as one affecting protein turn-over would plausibly affect the result. As noted above, Villers et al. (2012) used older mice, and their experiments were performed at a lower temperature than ours (28° vs. 31°) on slices that were not submerged but at interface. In addition, LTP was induced by a continuous high frequency stimulation (tetanus) protocol and not by the TBS stimulation used here. Critical time window for protein utilization The idea that triggered protein synthesis underlies LTP stabilization also rests on the finding that PSIs when applied some time after LTP induction does not impair LTP stability. That is, LTP must be stabilized by proteins synthesized in response to the LTP induction event. In an analogous manner, we found that while 4 h of CHX pre-TBS application partially de-stabilized LTP, >4 h post-TBS CHX application had no such effect. While a critical time window is easily understood if triggered proteins is important for LTP stabilization, it is however less so if constitutively synthesized proteins are involved. One possibility could be that the constitutively synthesized proteins critical for LTP stabilization instead have a critical time window for their utilization, instead of for their triggering. Another possibility could be that LTP induction results in posttranslational modifications of these proteins slowing their turnover. If so, the reduced level of these proteins following pre-LTP induction protein synthesis blockade would lead to a deficit of such proteins that would not appear following a postLTP induction protein synthesis blockade. ‘‘L-LTP’’ emerging after transient H2O2 application The numerous studies that have demonstrated a decaying LTP when a PSI has been applied at about the time of LTP induction still present an almost insurmountable challenge for the idea that blockade of constitutive rather than triggered protein synthesis underlies de-stabilization of LTP. However, as discussed by Villers et al. (2012), the stability of the LTP induced by a given induction protocol depends on experimental conditions such as the slice dissection procedures, slice recovery conditions, interface oxygenation and temperature control. It may be speculated that an improvement of these conditions may underlie the recent failure of these authors to affect LTP by PSI application (Villers et al., 2012 vs. Ris et al., 2009; Villers et al., 2010). In the present study we altered the experimental conditions by exposing the slices to oxidative stress by transiently applying H2O2 to the slices just after the LTP induction event. Such an application can result in damage of proteins, lipids, nucleic acids and other cellular components (Hunt et al., 1988; Gieseg et al., 1993; Shetty et al., 2008) as well as affect protein translation, increase the susceptibility of proteins to degradation by proteolytic enzymes (e.g. Fligiel et al., 1984), and enhance the removal of the
oxidized proteins by the proteasome system (Grune et al., 1995, 1996). In brief, this procedure ought to diminish the supply of constitutive proteins as well as possibly interfere with the production of new proteins. While this exposure resulted in a significant acute as well as long-term depression of the control input, after linearly adjusting for this depression, LTP of the potentiated input was as large and stable as normally observed. When performed in the presence of protein synthesis inhibition, however, LTP became destabilized. Taken at face value, these results suggest that a transient exposure of the slices to oxidative stress primes for the appearance of a protein turnoverdependent LTP stabilization. Considering the wide range of effects of H2O2, with often opposing actions at low and high concentrations, it is not possible to know what underlies the H2O2 effect on basal synaptic transmission or, in the presence of protein synthesis inhibition, on LTP. Nonetheless, the fact that H2O2 application in the middle of the critical period for triggered protein synthesis affects LTP no more than it does basal transmission seems to argue against this time period being critical for events important for LTP stabilization. Using a [3H]leucine assay we found that the H2O2 application resulted in a small but significant decrease in labeled protein, indicating a net decrease in cellular protein levels. However, there could be few pitfalls in this approach, most importantly is that only global rather than specific protein was estimated rendering unlikely to rule out the presence of highly degraded amounts of proteins that were obscured by the total estimated proteins. Furthermore, our choice of leucine, a commonly used radiolabeled amino acid, may confound the decay of proteins as it has been reported that tritiated leucine is reutilized for another turn of protein synthesis (e.g. Forgue and Dahl, 1978). Altogether, what is proposed here is that the depletion in protein repertoire to a degree that de-stabilizes LTP cannot be achieved before PSI application interval is long enough beyond the half-life of a set of proteins that is necessary for LTP stabilization. Given that short time interval of PSI application has not yielded an effect on LTP but such effect was obvious with longer treatment regime, it could be assumed that the half-life of such set is longer than 30 min, in our hand. However, when short treatment duration of CHX was associated with an acceleration of protein turn-over rate resulting from peroxide-mediated protein modification, LTP became more vulnerable. This is likely to be explained by CHX effect of curtailing the replenishment processes of those degraded proteins. However, presently, this scenario can only be seen as a working hypothesis. Whether the sensitivity of LTP for short time PSI protocols, as reported in the literature, might be explained by changing the kinetic of proteins, i.e. shortening half-life, remains to be systematically investigated. Acknowledgments—I would like to thank Dr. Jo¨rgen Ekstro¨m for providing facilities for tritiated leucine incorporation assay; and Ann-Christine Reinhold for excellent technical assistance. This study was supported by the LUA/ALF-agreement (ALFGBG-
A.-K. Abbas / Neuroscience 246 (2013) 301–311 11907) and the foundations of Wilhelm & Martina Lundgren, Lennanders, and Handlanden Hjalmar Svensson; the Swedish Medical Society and the Gothenburg Medical Society.
REFERENCES Abbas A-K, Dozmorov M, Li R, Huang F-S, Hellberg F, Danielson J, Tian Y, Ekstro¨m J, Sandberg M, Wigstro¨m H (2009) Persistent LTP without triggered protein synthesis. Neurosci Res 63:59–65. Abbas A-K, Huang F-S, Li R, Ekstro¨m J, Wigstro¨m H (2011) Emetine treatment masks initial LTP without affecting long-term stability. Brain Res 1426:18–29. Abraham WC, Williams JM (2008) LTP maintenance and its protein synthesis-dependence. Neurobiol Learn Mem 89:260–268. Dozmorov M, Niu Y-P, Xu H-P, Xiao M-Y, Li R, Sandberg M, Wigstro¨m H (2003) Active decay of composite excitatory postsynaptic potentials in hippocampal slices from young rats. Brain Res 973:44–55. Engert F, Bonhoeffer T (1999) Dendritic spine changes associated with hippocampal long-term synaptic plasticity. Nature 399:66–70. Fifkova E, Anderson CL, Young SJ, VanHarreveld A (1982) Effect of anisomycin of stimulation-induced changes in dendritic spines of the dentate granule cells. J Neurocytol 11:183–210. Fligiel SEG, Lee EC, McCoy JP, Johnson KJ, Varani J (1984) Protein degradation following treatment with hydrogen peroxide. Am J Pathol 115:418–425. Fonseca R, Na¨gerl UV, Morris RGM, Bonhoeffer T (2004) Competing for memory: hippocampal LTP under regimes of reduced protein synthesis. Neuron 44:1011–1020. Fonseca R, Vabulas RM, Hartl FU, Bonhoeffer T, Na¨gerl UV (2006) A balance of protein synthesis and proteasome-dependent degradation determines the maintenance of LTP. Neuron 52:239–245. Forgue ST, Dahl JL (1978) The turnover rate of tubulin in rat brain. J Neurochem 31:1289–1297. Frey U, Morris RGM (1997) Synaptic tagging and long-term potentiation. Nature 385:533–536. Frey U, Morris RGM (1998) Weak before strong: dissociating synaptic tagging and plasticity-factor accounts of late-LTP. Neuropharmacology 37:545–552. Gieseg SP, Simpson JA, Charlton TS, Duncan MW, Dean RT (1993) Protein-bound 3,4-dihydroxyphenylalanine is a major reductant formed during hydroxyl radical damage to proteins. Biochemistry 32:4780–4786. Grune T, Reinheckel T, Joshi M, Davies JA (1995) Proteolysis in cultured liver epithelial cells during oxidative stress. Role of the multicatalytic proteinase complex, proteasome. J Biol Chem 270:2344–2351.
311
Grune T, Reinheckel T, Davies JA (1996) Degradation of oxidized proteins in K562 human hematopoietic cells by proteasome. J Biol Chem 271:15504–15509. Hesse GW, Hofstein R, Shashoua VE (1984) Protein release from hippocampus in vitro. Brain Res 305:61–66. Hunt JV, Simpson JA, Dean RT (1988) Hydroperoxide-mediated fragmentation of proteins. Biochem J 250:87–93. Larson J, Wong D, Lynch G (1986) Patterned stimulation at the theta frequency is optimal for the induction of hippocampal long-term potentiation. Brain Res 368:347–350. Lipton P, Heimbach CJ (1978) Mechanisms of extracellular potassium stimulation of protein synthesis in guinea pig hippocampal slices. J Neurochem 31:1299–1307. Lynch MA (2004) Long-term potentiation and memory. Physiol Rev 84:87–136. Maletic-Savatic M, Malinow R, Svoboda K (1999) Rapid dendritic morphogenesis in CA1 hippocampal dendrites induced by synaptic activity. Science 283:1923–1927. Martin SJ, Grimwood PD, Morris RGM (2000) Synaptic plasticity and memory: an evaluation of the hypothesis. Annu Rev Neurosci 23:649–711. Matsuzaki M, Honkura N, Ellis-Davies GCR, Kasai H (2004) Structural basis of long-term potentiation in single dendritic spines. Nature 429:761–766. Ris L, Villers A, Godaux E (2009) Synaptic capture-mediated longlasting long-term potentiation is strongly dependent on mRNA translation. NeuroReport 20:1572–1576. Sajikumar S, Frey JU (2004) Late-associativity, synaptic tagging, and the role of dopamine during LTP and LTD. Neurobiol Learn Mem 82:12–25. Shetty PK, Huang FL, Huang K-P (2008) Ischemia-elicited oxidative modulation of Ca2+/calmodulin-dependent protein kinase II. J Biol Chem 283:5389–5401. Toni N, Buchs P-A, Nikonenko I, Bron CR, Muller D (1999) LTP promotes formation of multiple spine synapses between a single axon terminal and a dendrite. Nature 402:421–425. Tsuriel S, Geva R, Zamorano P, Dresbach T, Boeckers T, Gundelfinger ED, Garner CC, Ziv NE (2006) Local sharing as a predominant determinant of synaptic matrix molecular dynamics. PLoS Biol 4:1572–1587. Villers A, Godaux E, Ris L (2010) Late phase of L-LTP elicited in isolated CA1 dendrites cannot be transferred by synaptic capture. NeuroReport 21:210–215. Villers A, Godaux E, Ris L (2012) Long-lasting LTP requires neither repeated trains for its induction nor protein synthesis for its development. PLoS One 7:e40823. Volianskis A, Jensen MS (2003) Transient and sustained types of long-term potentiation in the CA1 area of the rat hippocampus. J Physiol 550:459–492.
(Accepted 2 May 2013) (Available online 16 May 2013)
EVIDENCE FOR CONSTITUTIVE PROTEIN SYNTHESIS IN HIPPOCAMPAL LTP STABILIZATION A.-K. ABBAS *
inhibitors (PSIs) during a critical time window (often ±30 min) around the LTP induction event makes an LTP, that would otherwise be long-lasting, to vanish within few hours, i.e., prevents late-phase LTP. Thus, an LTP-inducing tetanization can result in late-phase LTP only if this tetanization triggers synthesis of proteins stabilizing LTP. However, results from other studies have questioned this idea. For example, this effect of PSIs was not observed when also protein degradation was blocked indicating that constitutively expressed proteins may suffice for LTP stabilization (Fonseca et al., 2006). We and others have also recently failed to observe any effect of protein synthesis inhibition on LTP during its first 7–8 h (Abbas et al., 2009; Villers et al., 2012). Moreover, the stability of LTP following a given induction protocol may depend critically on experimental conditions, such as slice preparation and recovery conditions (for references, see Villers et al., 2012). That the state of the slice may affect LTP raises the question to what extent procedures affecting the protein content and/or the state of the proteins in the slice may affect LTP stabilization. Thus, can protein depletion per se result in an impaired ability of the synapses to undergo the structural/functional rearrangements necessary for stable LTP? In the present study we have introduced a long (4 h) pre-application of PSI to examine whether this procedure may result in less stable LTP. In another set of experiments we applied a mixture of anisomycin and cycloheximide (CHX) for 30 min before LTP induction throughout the time-course of recorded LTP. Furthermore, we also transiently applied hydrogen peroxide (H2O2) to the slice in order to induce rapid protein modification/degradation. Our findings suggest that proteins synthesized in response to the LTP-inducing stimulation are unlikely to be critical for the subsequent stabilization of LTP.
Institute of Neuroscience and Physiology, University of Gothenburg, P.O. Box 433, SE-40530 Gothenburg, Sweden
Abstract—The notion that blockade of constitutive protein synthesis underlies the effect of protein synthesis inhibitors (PSIs) on long-term potentiation (LTP) stabilization was examined using the rat hippocampal CA3–CA1 synapse. Using a biochemical assay we found protein synthesis rate largely recovered 1 h after wash-out of cycloheximide (CHX). Nonetheless, a 4-h CHX application followed by wash-out 1 h prior to LTP resulted in a significant decrement of LTP stabilization. Wash-out initiated just prior to LTP, thus extending protein synthesis inhibition well into the post-LTP period, resulted in no further effect on LTP. However, short pre- and continuous post-tetanization application of PSIs failed to influence LTP persistence for up to 7 h. Addition of hydrogen peroxide (H2O2) 5–25 min following LTP induction resulted in parallel depression of potentiated and non-potentiated inputs, leaving LTP seemingly unaltered. However, in the presence of cyxloheximide the H2O2 application resulted in a significant reduction of LTP. In conclusion: LTP stabilization was impaired by pre-LTP application of protein synthesis inhibition but not by postLTP application unless the slices were exposed to oxidative stress. We submit that these results favor the notion that constitutive rather than triggered protein synthesis is important for LTP stabilization. Ó 2013 IBRO. Published by Elsevier Ltd. All rights reserved.
Key words: anisomycin, cycloheximide, neuronal plasticity, oxidative stress, protein synthesis, synapse.
INTRODUCTION Long-term potentiation (LTP) is a cellular model for some forms of learning and memory (Martin et al., 2000; Lynch, 2004). In common with memory LTP has been differentiated temporally into an early vs. late phase, the latter believed to require the synthesis of novel proteins for its expression. In fact, a large number of studies have shown that application of protein synthesis
EXPERIMENTAL PROCEDURES Animals The experimental biomedicine (EBM) animal facility is fully accredited by the Swedish Council for Laboratory Animals. Fourteen- to 22-day-old Albino rats (strain Sprague–Dawley) of mixed sexes were obtained from Charles River (Scanbur AB, Sollentuna, Sweden) and were prior to the experiment kept group-housed. By using a multi-recording chamber system allowing four slices to be studied in parallel, the number of animals used was kept at a minimum.
*Tel: +46-768-235900; fax: +46-31-7863840. E-mail address: [email protected] Abbreviations: ACSF, artificial cerebrospinal fluid; AP-5, D-( )-2amino-5-phosphonopentanoic acid; CHX, cycloheximide; DMSO, dimethysulfoxide; DPCPX, 8-cyclopentyl-1,3-dipropylxanthine; fEPSP, field excitatory postsynaptic potentials; FeSO4, ferrous sulfate; H2O2, hydrogen peroxide; LTP, long-term potentiation; PSIs, protein synthesis inhibitors; TBS, theta-burst stimulation; TCA, trichloroacetic acid.
0306-4522/13 $36.00 Ó 2013 IBRO. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.neuroscience.2013.05.011 301
302
A.-K. Abbas / Neuroscience 246 (2013) 301–311
In vitro electrophysiology and drugs preparation The animals were killed in accordance with the guidelines of the local ethics committee of University of Gothenburg. They were deeply anesthetized with isoflurane (Baxter Medical AB, Oslo), decapitated and the brain was removed. Transverse hippocampal slices (400-lm-thick) were cut using a McIlwain-type tissue chopper and submerged in a holding chamber containing roomtempered artificial cerebrospinal fluid (ACSF) composed of (in mM): 119 NaCl, 2.5 KCl, 2.0 CaCl2, 2.0 MgCl2, 26 NaHCO3, 1.0 NaH2PO4, and 10 D-glucose, equilibrated with 95% O2 and 5% CO2. After pre-incubation for at least 90 min, slices were transferred as needed to one or several recording chambers where they were kept submerged and covered with nylon net to prevent their movement. The recording chamber consisted of a circular well of low volume (1.5–2 ml), and the solution was recycled (1.5–2.0 ml/min) using a peristaltic pump (Ismatec, Labinett Lab AB, Sweden). The perfusion solution was the same as in the holding chamber except for 2.5 mM CaCl2 and 1.3 mM MgCl2. Experiments were performed at 31 °C. Extracellular field potentials were recorded with a glass micropipette filled with 1 M NaCl (resistance 3–5 MX) positioned in the middle of the CA1 stratum radiatum. Microelectrodes were pulled from microfiber (o.d. 1.5 mm, i.d. 0.86 mm, Warner Instruments, LLC, Hamden, CT, USA) capillary tubing. To stimulate independent inputs to the same cell population, two monopolar tungsten stimulating electrodes (0.1 MX; World Precision Instrument, Inc., Sarasota, FL, USA) were positioned on either side of the recording microelectrode. Their positions were arranged so that the same amount of stimulating current evoked field potentials of similar magnitude. Stimuli were delivered to the commissural–Schaffer collateral afferents as 100-ls negative constant-current pulses (20–50 lA) using a programmable pulse generator. While one synaptic input (test pathway) was employed to induce and monitor LTP, the other one (control pathway) was used to verify the stability of basal synaptic transmission. Pulses were delivered alternately to the two pathways, usually at a rate of once every 40 or 60 s, with stimuli to the two pathways separated by 20 or 30 s, respectively. Baseline field excitatory postsynaptic potentials (fEPSPs) were recorded for 60–90 min to ensure stability of responses. LTP was induced generally by theta-burst stimulation (TBS) composed of three trains (inter-train interval 5 s), each consisting of 10 bursts of four pulses at 100 Hz, with an inter-burst interval of 200 ms (Larson et al., 1986). In one group of experiments, LTP was induced by three tetanus trains (100 Hz, 1 s, inter-train interval 5 s). The PSIs anisomycin (2-[p-methoxybenzyl]-3,4, pyrrolidinediol-3-acetate) and CHX 4-{(2R)-2[(1S,3S,5S)-3,5-dimethyl-2-ococyclohexyl]-2-hydroxyethyl} piperidine-2,6-dione, H2O2 (3%), ferrous sulfate (FeSO4) and dimethysulfoxide (DMSO) were all purchased from Sigma–Aldrich (St. Louis, MO, USA). 8-Cyclopentyl-1,3dipropylxanthine (DPCPX) and D-( )-2-amino-5phosphonopentanoic acid (AP5) were obtained from
Tocris Bioscience (UK) or Ascent Scientific Ltd (UK). [3H]leucine was obtained from Amersham, Buckinghamshire, UK. Milli-Q deionized water (Millipore, Bedford, MA, USA) was used in all preparations of buffers and solutions. Other chemicals used were all of highest grade commercially available. For biochemical assay, anisomycin, or CHX, was added to ACSF to different final concentrations derived from different stocks prepared in distilled water at the time of each experiment. For electrophysiological experiments, CHX and DPCPX were first dissolved in DMSO to a concentration of 100 and 3 mM, respectively, and stored at 18 °C until diluted in physiologic buffer to final concentration of 100 lM and 50 nM, respectively. The largest final concentration of DMSO was 0.1%, which had no effect on basal synaptic transmission as verified by lack of effect on the baseline fEPSP when added as vehicle. Stocks of AP5, anisomycin and FeSO4 were prepared in distilled water and diluted at final concentrations of 50 lM, 25 mM and 100 lM, respectively. The rational for FeSO4 addition was to enhance the generation of hydroxyl radical from H2O2 via a Fenton-like reaction. Biochemical assays Effect of PSIs on protein synthesis was measured by [3H]leucine incorporation (Lipton and Heimbach, 1978) in whole slices. Hippocampal slices from 13- to 24-dayold rats were maintained under similar conditions as in the electrophysiological work but without stimulation. Slices were put on a multiwall plastic dish (Corning Incorporated, Corning, NY, USA) and assigned to a PSI group or a control group in an interleaved manner to minimize inter-slice variability with respect to weight and metabolism. To test the effect of PSIs on rate of protein synthesis in dose-dependent manner, two slices per group were pre-incubated for 10 min with the drug (or vehicle, if necessary) before addition of [3H]leucine; final activity 0.5–1 lCi/ml. Uptake and incorporation of leucine into proteins were allowed to proceed for 50 min either with or without PSI. To test PSI reversibility, a corresponding total duration of pre-incubation with the drug (60 min) was followed by different pre-incubation intervals (30, 60 or 90 min) in drug-free solution before allowing slices to be incubated in a [3H]leucinecontaining solution for additional 50 min. Radioactive amino acid incorporation was then terminated by the addition of ice-cold saline, and the slices were subsequently lysed in 1 ml of 5 mM NaOH. After protein purification, incorporation of leucine into trichloroacetic acid (TCA)-precipitable proteins was measured in a scintillation counter (LKB Wallace, 1219 Rackbeta, Finland). Percentage inhibition of leucine incorporation produced by drug treatment was calculated by comparing counts in treated slices with those of control ones. To investigate whether protein content was influenced by the addition of H2O2, two groups of slices were incubated with tritiated leucine for 50 min at 31 °C following a pre-incubation period equivalent to that for the electrophysiological experiments. The slices were subsequently washed by continuous perfusion with
A.-K. Abbas / Neuroscience 246 (2013) 301–311
303
ACSF for 30 min before H2O2/FeSO4 was applied to one group of slices for 20 min at concentrations identical to those used for the electrophysiological experiments. Both control and H2O2/FeSO4-exposed slices were subsequently rinsed and frozen at different time intervals following the exposure. Percentage of leucine incorporation produced by H2O2/FeSO4 was calculated by comparing counts in treated slices with those of control slices. Data analysis Electrophysiological data were collected using in-house developed electronic equipment based on an Eagle Instruments multifunction board, capturing signals from up to four parallel experiments. Signals were amplified, filtered, digitized and transferred to a digital computer for on-line and off-line analysis. Off-line data analysis was performed using a combination of Quickbasic (Microsoft) and Igor (Wavemetrics, Inc.) programs. fEPSPs were measured using an early time window of 1–2-ms duration positioned just after the presynaptic volley. The area under the curve was used as a measure of fEPSP size. Alternatively, the fEPSP can be quantified by a slope measurement, shown to give similar results (Dozmorov et al., 2003). The area measurement was generally preferred in the present study due to less noise and better immunity to volleyEPSP overlap. Slices in which rundown of the control fEPSP was by more than 30% at the end of experiments (10 h of recording) were excluded (Volianskis and Jensen, 2003). For time course plots, LTP values were averaged over 7 min (7–10 successive responses depending on test frequency). For each successive measurement, the mean, standard deviation, and standard errors were calculated for all the experiments. fEPSP (mV/ms) was calculated and data for each experiment were normalized relative to baseline recording. As normality assessment (Q–Q plot and Shapiro–Will test) revealed non-Gaussian distribution of groups, nonparametric Mann–Whitney U test was conducted (SPSS for windows, Version 18.0) when data were compared between groups (drug vs. control). Alternatively, Student’s t-test was used to assess comparisons when the distribution of groups was Gaussian. Numbers of experiments are indicated by n. Values of p < 0.05 were considered to represent significant differences.
RESULTS Effect of prolonged post-induction application of PSIs on LTP We reported previously (Abbas et al., 2009) that anisomycin or emetine applied from 30 min prior to 30 min after the LTP induction event, thus covering the proposed critical interval of triggered protein synthesis (Frey and Morris, 1997, 1998), failed to affect LTP recorded for up to 8 h. As shown in Fig. 1, we also observed such failure when wash-out was omitted, the tetanization resulting in seemingly equally stable LTP in
Fig. 1. Lack of effect on LTP by continuous application of PSIs. LTP was induced by three trains of tetanus (100 Hz, 1- and 5-s inter-train interval) in one of two separate pathways, using the other pathway as control. (A) LTP induced in non-treated slices. (B) LTP induced in slices in the presence of anisomycin (25 lM) and cycloheximide (70 lM) as indicated by the horizontal bar. (C) LTP shown in A and B are plotted superimposed after drift-compensation (test/control ratio) indicating no effect of the PSI application. Insets in A and B show superimposed fEPSPs of control and test pathways taken before, 4 h after, and 7 h after induction of LTP. Arrows indicate time of LTP induction. Each value in the graph represents the average fEPSP (±S.E.M.) sampled over 10-min periods.
the treated group of slices (Fig. 1B, n = 5) as in the non-treated group (Fig. 1A, n = 7). It can be noted that the drift of the control pathways was small and did not differ between the two groups suggesting that the prolonged protein synthesis inhibition by itself did not affect the synaptic transmission. To directly compare the LTP we compensated for the drift by computing the ratio of test to control responses. As shown in Fig. 1C, where
304
A.-K. Abbas / Neuroscience 246 (2013) 301–311
these drift-compensated LTP are shown superimposed, they were closely overlapping. Thus, in contrast to Ris et al. (2009); but see Villers et al. (2012), we failed to observe that a prolongation of the post-induction drug application results in a more efficient blockade of LTP. In the above experiments, two PSIs were used together (anisomycin, 20 lM and CHX, 80 lM), avoiding emetine which has nonspecific effects when applied for a long duration (Abbas et al., 2011). Nonetheless, in view of our negative results we realized the need to control for the efficacy of these drugs under conditions that faithfully match those of our present electrophysiological studies. Moreover, since in the remaining part of our study the protein synthesis blockade should occur only prior to the LTP induction event, we needed to establish the time course of reversal from the protein synthesis blockade following drug wash-out. Determination of potency of the PSIs Fig. 2A shows percentage inhibition of protein labeling using 50-min incubation with [3H]leucine and a variable concentration of anisomycin, indicating an IC50 value in the order of 50 nM. Even the lowest concentration, 10 nM, resulted in a measurable inhibition (about 20%) and a saturating inhibition (>90%) was achieved with concentrations P 5 lM. Doses in 10–25 lM range are commonly used to inhibit protein synthesis and were previously employed by us (Abbas et al., 2009, 2011). As evidenced by the results in Fig. 2A, such concentrations of anisomycin should be capable of efficient inhibition of protein synthesis. Fig. 2B shows the inhibition profile obtained for CHX, indicating an IC50 value of 300 nM, substantially higher than the corresponding value for anisomycin. Saturation (95% inhibition) was reached by concentrations of P40 lM. To ascertain that the mixture of CHX and anisomycin in the above experiments resulted in an optimal protein synthesis inhibitory effect, we conducted a [3H]leucine incorporation assay. As shown in Fig. 2C, the mixture of anisomycin and CHX resulted in 95 ± 0.4% (n = 4) protein synthesis inhibition. Notably, the level of protein synthesis inhibition produced by either anisomycin (Fig. 2A) or CHX (Fig. 2B) alone was not significantly different from that produced by the mixture (anisomycin, 93 ± 1%, n = 8; CHX, 94 ± 0.6%, n = 8). Reversibility of protein synthesis blockade following drug wash-out [3H]leucine assays were used to test the reversibility of anisomycin- and CHX-induced protein synthesis blockade. The inhibitor was first applied for 60 min, this application followed by wash-out and variable intervals (0–90 min) before [3H]leucine (without PSI) was applied for a 50-min period (Fig. 3). The experiment was finalized by analyzing protein content (see Experimental procedures, and the illustrated scheme in Fig. 3A). The histogram data in Fig. 3B shows that both anisomycin and CHX at 0-min wash-out essentially fully blocked
Fig. 2. Dose-dependent effect of PSIs on [3H]leucine incorporation into hippocampal slices. (A) Slices, at each concentration of anisomycin used, were first pre-incubated either with normal solution (control) or with anisomycin (test) for 10 min before adding radioactive leucine to these solutions for further 50 min. (B) Same as in A but using cycloheximide. The n values (equal for test and control at a given concentration) ranged between four and eight for anisomycin experiments, and between four and 12 for the cycloheximide experiments (part of these data has been published in Abbas et al., 2011). (C) Maximal level of protein synthesis inhibition (±S.E.M.) obtained by cycloheximide (CHX; n = 8), anisomycin (Ani; n = 7), and a mixture of Ani plus CHX (n = 4), indicating no significant difference between these treatments.
protein synthesis (as was the case with overlapping PSI and [3H]leucine application shown above). When longer wash-out periods (30, 60, and 90 min) were allowed prior to [3H]leucine application the CHX-induced inhibition decreased substantially. For instance, after 30 and 60 min of wash-out, protein synthesis rate was up to 68% and 80%, respectively, of control. The protein synthesis inhibition induced by anisomycin was less reversible, the protein synthesis rate no more than 10% of control even after 90 min of wash-out.
Effect of prolonged pre-TBS application of CHX, with pre-TBS washout, on LTP In view of the poor reversibility of the anisomycin-induced blockade, only CHX was used to test the effect of an extended period of pre-TBS protein synthesis inhibition on LTP. In the design of these experiments the period of drug application ought to be long enough to cause a significant depletion of a wide range of constitutively
A.-K. Abbas / Neuroscience 246 (2013) 301–311
305
normal and drug groups, respectively; U = 5; r = 0.6; p < 0.05; Mann–Whitney U test). Effect of prolonged pre-TBS application of CHX without pre-TBS washout on LTP
Fig. 3. Time-course of PSI reversibility. (A) Schematic illustration of the experiments. Anisomycin or cycloheximide was first applied for 60 min, followed by a variable washout interval before [3H]leucine assays was conducted for further 50 min. The experiment was eventually finalized by analyzing protein content. Four time points were estimated with the one with 0 wash-out was used as control (a). Values from three washout intervals: 30 min (b), 60 min (c) and 90 min (d) were compared with the control group (a). (B) Assessing time-course of reversibility of anisomycin (25 lM; n = 4 for each time point; black bars) or cycloheximide (40 lM; n = 4 for each time point; gray bars) using tritiated leucine incorporation showed that while washout interval following cycloheximide treatment requires >30 min for protein synthesis recovery to reach 80%, anisomycin washout was not followed by protein synthesis recovery up to 90 min. The group mean percentage (±S.E.M) for each time point of washout following protein synthesis inhibition computed for n slices in the protein synthesis inhibitor group and n slices in the control group without washout interval (dashed line; 100%).
synthesized proteins. Given that proteins involved in synaptic plasticity may have half-lives up to 2–3 h in vitro (Sajikumar and Frey, 2004), that mammalian extracellular proteins have half-lives in the order of 4 h (Hesse et al., 1984), and that tag lifetime has been reported to be 1–2 h (Frey and Morris, 1997, 1998; but see Fonseca et al., 2004), we chose a 4-h CHX application paradigm, followed by 1 h of wash-out before LTP induction and subsequent 7 h of LTP recording. This paradigm was experimentally feasible since total recording time in our hands can generally extend up to 12–13 h. These results obtained using CHX was compared with those obtained using non-treated slices in which the LTP-inducing TBS trains were given following 5 h of baseline stimulation. Fig. 4 shows LTP obtained in non-treated (A) and treated (B) slices, respectively, without compensation for control input drift, and the corresponding drift-compensated LTP are shown superimposed in (C). It can be noted that while LTP values of non-treated and treated slices did not differ when measured 1 h after induction (95 ± 9% vs. 91 ± 15%; U = 16; r = 0.09; p > 0.05; Mann–Whitney U test) they did significantly differ 5 h later (47 ± 13% vs. 10 ± 9%; U = 4; r = 0.65; p < 0.05; Mann–Whitney U test). The drift-compensated LTP revealed a significant difference at 8 h (60 ± 5% vs. 40 ± 4%, for
In initial experiments prior to the realization of the slow reversibility of the protein synthesis inhibition, LTP was induced soon after the onset of CHX wash-out. Thus, in these experiments protein synthesis was not only blocked for 4 h preceding LTP, but to a substantial extent also for roughly 1 h post-TBS, thus covering much of the critical period believed to exist for the triggered protein synthesis. Fig. 5 shows that also under this condition the LTP of non-treated (A) and treated (B) slices did not differ when evaluated 1 h after induction (110 ± 8% (n = 8) vs. 114 ± 11% (n = 14); U = 42; r = 0.2; p > 0.05; Mann–Whitney U test) but differed significantly 5 h later (63 ± 10% vs. 35 ± 12%; U = 22; r = 0.5; p < 0.05; Mann–Whitney U test). Values obtained as test/control input ratio (Fig. 5C) at 6 h were in order of 190 ± 7% vs. 161 ± 6% (U = 15; r = 0.6; p < 0.005; Mann–Whitney U test). The parallel findings evident both from the original data in Figs. 5A, B and the drift-compensated ones in Fig. 5C were a consequence of quantitatively similar rundown of the control pathway among the two groups. Effect of H2O2 application on synaptic transmission and on LTP To induce protein modification/degradation following LTP induction H2O2 (5 mM) together with FeSO4 (100 lM) was applied for 20 min, starting 5 min following the LTP induction event. To reduce some unwanted side effects, the experiments were performed in the presence of the adenosine A1 receptor antagonist DPCPX (50 nM), and the NMDA receptor antagonist AP5 (50 lM) was added after the LTP induction. Fig. 6A shows that H2O2/FeSO4 application had a substantial immediate depressing effect on the control inputs (n = 6) that remained to much the same extent after H2O2/FeSO4 wash-out and throughout the 7 h post-LTP induction recording period. This depressive effect on the control inputs was seemingly mirrored by a similar effect on the LTP of the potentiated inputs (in order of 141 ± 13% and 126 ± 14% at 2 and 7 h, respectively; Fig. 6B). In Fig. 6C LTP exhibited as ‘‘drift-compensated’’ is shown to be stable indicating a parallel depression to that of the control input. Measurements of the presynaptic fiber volley indicated no effects of H2O2/FeSO4 application on axon excitability (data not shown). CHX reduces LTP H2O2/FeSO4 was subsequently applied to slices in the presence of CHX (80–100 lM), whose application started 30 min before LTP induction and kept throughout the 7-h recording period. As shown in Fig. 7A, B both control and potentiated inputs (n = 11) became depressed by the H2O2/FeSO4 application in much the same manner as in the absence of CHX. However, in
306
A.-K. Abbas / Neuroscience 246 (2013) 301–311
Fig. 4. Effect of long pre-incubation with cycloheximide prior to LTP induction. (A) LTP was induced by a TBS paradigm (three trains of 10 bursts of four pulses at 100 Hz, repeated with a burst frequency 5 Hz, 5-s inter-train interval) in one of two separate pathways, using the other as control (n = 6). TBS trains were given following 5 h of baseline stimulation. (B) Same as in (A), but in slices that had been pre-incubated with cycloheximide (100 lM) for 4 h, the cycloheximide wash-out beginning 1 h prior to LTP induction (n = 6). (C) LTP shown in (A, B) is plotted superimposed after drift-compensation (test/control ratio) indicating less stability of LTP in PSI-treated slices. Insets in A and B show superimposed fEPSPs of control and test pathways taken before, 4 h after, and 7 h after induction of LTP. Arrows indicate time of LTP induction. Each value in the graph represents the average initial area of fEPSPs (±S.E.M.) sampled over 8 min periods.
the presence of CHX LTP appears disproportionally affected (in order of 121 ± 12% and 90 ± 8% at 2 and 7 h, respectively; Fig. 7B). We thus computed the ‘‘driftcompensated’’ LTP by taking the ratio test/control responses to assess the effect of CHX on LTP itself. In Fig. 7C these ‘‘drift-compensated’’ LTP are shown superimposed, indicating little effect of protein synthesis inhibition during the first 3–4 h of LTP. However, thereafter, LTP in CHX-treated slices decays substantially implying a role for protein breakdown in LTP stabilization following the transient H2O2/FeSO4
application. Significant LTP values were observed at 7 h following LTP induction amounting 195 ± 16% vs. 145 ± 8% (t = 3.2078; df = 15; p < 0.05; Student’s t test) for H2O2/FeSO4 and H2O2/FeSO4/CHX group, respectively. H2O2 causes a significant reduction in the newly synthesized proteins We also examined the effect of H2O2 (5 mM, 20 min) on the turnover of newly synthesized proteins. The
A.-K. Abbas / Neuroscience 246 (2013) 301–311
307
Fig. 5. Effect of long pre-incubation with cycloheximide prior to LTP induction. (A) LTP was induced by a TBS paradigm (ten 4-impulse 100 Hz trains at 5 Hz, repeated thrice at 0.2 Hz) in one of two separate pathways, using the other as control (n = 6). TBS trains were given following 4 h of baseline stimulation. (B) Same as in (A), but in slices that had been pre-incubated with cycloheximide (100 lM) for 4 h, the cycloheximide wash-out beginning immediately prior to LTP induction. (C) LTP shown in (A, B) is plotted superimposed after drift-compensation (test/control ratio) indicating less stability of LTP in PSI-treated slices. Insets in (A, B) show superimposed fEPSPs of control and test pathways taken before, 4 h after, and 7 h after induction of LTP. Arrows indicate time of LTP induction. Each value in the graph represents the average initial area of fEPSPs (±S.E.M.) sampled over 8 min periods.
histogram data in Fig. 8 show that the percentage of newly synthesized radiolabeled proteins 1 h following H2O2 washout was nearly stable compared to control values (100%). However, at 3 h following washout it decreased to moderate but significant degree amounting 85 ± 4% (p < 0.05; cf. Grune et al., 1995, 1996).
DISCUSSION Three main results emerge from the present study. First, that an extended pre-TBS period of protein
synthesis blockade destabilizes LTP maintenance simulating a partial blockade of ‘‘L-LTP’’. Second, continuous protein synthesis application after LTP induction did not cause an observable effect on LTP stabilization. Third, that after a transient post-TBS application of H2O2, but not in control conditions, protein synthesis inhibition covering the ‘‘critical’’ time window (first hour post-TBS) could destabilize LTP. As will be detailed below, we believe these results favor the idea that the supply of constitutively expressed, rather than of tetanization-triggered,
308
A.-K. Abbas / Neuroscience 246 (2013) 301–311
Fig. 6. Effect of oxidative stress on basal transmission and on LTP. Slices were in the presence of DPCPX and AP5 exposed to a 20 min application of hydrogen peroxide (H2O2) plus FeSO4. The rational for adding FeSO4 was to enhance hydroxyl radical formation via a Fenton-like reaction. (A) Control pathway. Note the immediate and lasting depression induced by this application. (B) Test pathway activated by LTP-inducing TBS (ten 4-impulse 100 Hz trains at 5 Hz, repeated thrice at 0.2 Hz) 5 min prior to the hydrogen peroxide (H2O2) plus FeSO4 application. Each value in the graph represents the average initial area of fEPSPs (±S.E.M.) sampled over 5-min periods. Insets in (A, B) show superimposed field EPSPs taken at times indicated in the graphs. (C) The ‘‘drift-compensated’’ LTP (test/ control ratio). Note that the magnitude and stability of LTP are as much as normally observed (cf. Fig. 5C) implying that the effect of hydrogen peroxide on either input is due to a reduction in the basal synaptic transmission.
proteins determines the stability of LTP maintenance at least within its first 8 h. Potency and time-course of protein synthesis inhibition For the interpretation of the present results it was vital that the pre-TBS protein synthesis blockade was largely restricted to the pre-TBS time period. At the
Fig. 7. Effect of cycloheximide (CHX) on basal transmission and on LTP in slices exposed to oxidative stress. The results shown in Fig. 6 are re-plotted in this figure (open circles) together with results obtained in slices had been exposed to cycloheximide (100 lM) throughout the experiment (starting 30 min prior to LTP induction) (closed circles). This set of experiments is also performed in the presence of DPCPX and AP5 at similar concentrations and time points. Note that while the field EPSPs of the control pathway (A) are little affected by the presence of cycloheximide, the field EPSPs of the tetanized pathway (B) are more affected, especially in the later part of the LTP. Insets in (A, B) illustrate superimposed fEPSP taken at times indicated in the graphs. In (C), where the drift-compensated LTP observed in the presence of cycloheximide is shown superimposed with that obtained in the absence of this drug, shows that cycloheximide impairs LTP stability in slices exposed to oxidative stress.
concentrations presently used, anisomycin and CHX were found to inhibit protein synthesis by 90–95%, but how reversible was this inhibition following drug washout? Here we found, using a [3H]leucine assay, that while both drugs had fast onset times (a few minutes), in the case of anisomycin no reduction of the inhibitory
A.-K. Abbas / Neuroscience 246 (2013) 301–311
309
Constitutive proteins in LTP stabilization
Fig. 8. Time course of protein labels percentage in hippocampal slice following [3H]leucine and H2O2/FeSO4. Pooled values (±S.E.M.) from eight and four experiments, for 1- and 3-h time points, respectively, compared to their own normalized control slices (broken line). The data reveal a significant decrease in the labeled proteins at 3 h but not 1 h following peroxide treatment. Each experiment is average of two slices.
effect was observed even after nearly 2 h of washout. This result agrees with that of Villers et al. (2012) who showed that anisomycin even after 4 h of wash-out inhibited protein synthesis rate by more than 80%. On the other hand, the CHX effect was reversible but still required about 1 h of wash-out for the synthesis rate to rise to 80% of control. In our pre-TBS experiments CHX was used with wash-out starting 1 h prior to the LTP induction, implying that some degree of protein synthesis inhibition would still extend into the ‘‘critical’’ time window for de novo synthesis of proteins believed to stabilize LTP. However, since in separate experiments in which the wash-out was started just before LTP induction instead of 1 h before did not result in a less stable LTP, it seems safe to conclude that the effect of our pre-TBS application is related to a pre-TBS rather than to a post-TBS protein synthesis blockade.
No role for triggered protein synthesis In our previous work (Abbas et al., 2009), application of neither emetine nor anisomycin for ±30 min around the time of LTP induction was found to influence LTP maintenance. In consideration of the possibility that the period of triggered de novo protein synthesis critical for LTP maintenance might vary among different studies depending on experimental conditions, in initial experiments we extended the post-tetanization protein synthesis blockade by omitting drug wash-out. We also applied a cocktail of anisomycin and CHX to maximize the inhibitory effect on protein synthesis. Despite these efforts, which considering our later observed high level of inhibition by each of these drugs alone and poor reversibility of the anisomycin-induced inhibition, may be found somewhat exaggerated, no de-stabilization of LTP maintenance was observed. Thus, under our experimental conditions triggered protein synthesis is not critical for the stabilization of LTP at least within its first 8 h. This lack of effect of prolonged posttetanization protein synthesis blockade was recently also reported by Villers et al. (2012) using older rats (6– 10 weeks), indicating that our use of young rats (2– 3 weeks) has not been decisive for obtaining this result.
The idea that LTP is stabilized beyond its first few hours by proteins whose synthesis is triggered by the LTP-inducing high-frequency stimulation rests almost entirely on the observation from numerous studies that application of PSIs during the ‘‘critical’’ time window results in an LTP that is largely decayed within few hours. The nature of these proteins and in what manner they should stabilize LTP are questions essentially unanswered (see e.g. Discussion in Villers et al., 2012). In general terms these proteins should ‘‘cement the synaptic change initiated during early phase LTP and contribute to activitydependent structural changes’’ (Fifkova et al., 1982; described by Abraham and Williams, 2008). In fact, the findings reported by Fifkova et al. (1982) were problematic in term that early (4 min) but not late (90 min) entorhinal spine morphological changes were prevented by administration of anisomycin in vivo. However, neither an early phase LTP expression based on e.g. CaMKII autophosphorylation nor activity-dependent structural changes will necessarily need new proteins. For example, rapid morphological changes such as persistent spine enlargement were observed within less than 1 min following glutamate receptors stimulation (Matsuzaki et al., 2004). Furthermore, emergence of new spines was documented to occur 30–60 min following high-frequency stimulation (Maletic-Savatic et al., 1999; Toni et al., 1999) or LTP induction (Engert and Bonhoeffer, 1999). Although there are no reports, up to our knowledge, confirming whether late morphological changes associated with LTP in vitro might be achieved without the requirement for triggered de novo protein synthesis, dendritic and/or neighboring macromolecules redistribution might not be ruled out. In this context, Tsuriel et al. (2006) showed that in cultured rat hippocampal neurons the dynamics of molecules, such as synapsin and ProSAP2, at individual CNS synapses may be dominated primarily by the continuous exchange and redistribution among nearby synapses lasting for up to 6 h, whereas protein synthesis and degradation may constitute slower, second-order processes that serve to maintain and regulate the size of local, shared pools of synaptic matrix proteins. Additional data confirmed that these dynamics are not unique to synapsin I and ProSAP2 but are characteristic of synaptic cytomatrix protein in general (for references, see Tsuriel et al., 2006). Some studies have also previously shown that LTP, albeit under specific experimental circumstances, can be stable without a tetanization-triggered synthesis of new proteins. Thus, if test pulse stimulation is interrupted post-tetanus or if protein degradation is simultaneously inhibited, LTP is maintained. Such results would then rather point to the necessity of a sufficient supply of proteins synthesized prior to LTP induction for stabilization of LTP, at least within its first 8 h. Our partial de-stabilization of LTP by a prolonged pre-TBS protein synthesis blockade would then concur with such an interpretation of the protein synthesis dependence of LTP. Villers et al. (2012) did, in contrast to us, not find any effect of a 4-h pre-application of anisomycin on LTP recorded for up to 8 h. It should
310
A.-K. Abbas / Neuroscience 246 (2013) 301–311
then first be noted that we observed but a partial de-stabilization of LTP, which implies that any difference in experimental conditions such as one affecting protein turn-over would plausibly affect the result. As noted above, Villers et al. (2012) used older mice, and their experiments were performed at a lower temperature than ours (28° vs. 31°) on slices that were not submerged but at interface. In addition, LTP was induced by a continuous high frequency stimulation (tetanus) protocol and not by the TBS stimulation used here. Critical time window for protein utilization The idea that triggered protein synthesis underlies LTP stabilization also rests on the finding that PSIs when applied some time after LTP induction does not impair LTP stability. That is, LTP must be stabilized by proteins synthesized in response to the LTP induction event. In an analogous manner, we found that while 4 h of CHX pre-TBS application partially de-stabilized LTP, >4 h post-TBS CHX application had no such effect. While a critical time window is easily understood if triggered proteins is important for LTP stabilization, it is however less so if constitutively synthesized proteins are involved. One possibility could be that the constitutively synthesized proteins critical for LTP stabilization instead have a critical time window for their utilization, instead of for their triggering. Another possibility could be that LTP induction results in posttranslational modifications of these proteins slowing their turnover. If so, the reduced level of these proteins following pre-LTP induction protein synthesis blockade would lead to a deficit of such proteins that would not appear following a postLTP induction protein synthesis blockade. ‘‘L-LTP’’ emerging after transient H2O2 application The numerous studies that have demonstrated a decaying LTP when a PSI has been applied at about the time of LTP induction still present an almost insurmountable challenge for the idea that blockade of constitutive rather than triggered protein synthesis underlies de-stabilization of LTP. However, as discussed by Villers et al. (2012), the stability of the LTP induced by a given induction protocol depends on experimental conditions such as the slice dissection procedures, slice recovery conditions, interface oxygenation and temperature control. It may be speculated that an improvement of these conditions may underlie the recent failure of these authors to affect LTP by PSI application (Villers et al., 2012 vs. Ris et al., 2009; Villers et al., 2010). In the present study we altered the experimental conditions by exposing the slices to oxidative stress by transiently applying H2O2 to the slices just after the LTP induction event. Such an application can result in damage of proteins, lipids, nucleic acids and other cellular components (Hunt et al., 1988; Gieseg et al., 1993; Shetty et al., 2008) as well as affect protein translation, increase the susceptibility of proteins to degradation by proteolytic enzymes (e.g. Fligiel et al., 1984), and enhance the removal of the
oxidized proteins by the proteasome system (Grune et al., 1995, 1996). In brief, this procedure ought to diminish the supply of constitutive proteins as well as possibly interfere with the production of new proteins. While this exposure resulted in a significant acute as well as long-term depression of the control input, after linearly adjusting for this depression, LTP of the potentiated input was as large and stable as normally observed. When performed in the presence of protein synthesis inhibition, however, LTP became destabilized. Taken at face value, these results suggest that a transient exposure of the slices to oxidative stress primes for the appearance of a protein turnoverdependent LTP stabilization. Considering the wide range of effects of H2O2, with often opposing actions at low and high concentrations, it is not possible to know what underlies the H2O2 effect on basal synaptic transmission or, in the presence of protein synthesis inhibition, on LTP. Nonetheless, the fact that H2O2 application in the middle of the critical period for triggered protein synthesis affects LTP no more than it does basal transmission seems to argue against this time period being critical for events important for LTP stabilization. Using a [3H]leucine assay we found that the H2O2 application resulted in a small but significant decrease in labeled protein, indicating a net decrease in cellular protein levels. However, there could be few pitfalls in this approach, most importantly is that only global rather than specific protein was estimated rendering unlikely to rule out the presence of highly degraded amounts of proteins that were obscured by the total estimated proteins. Furthermore, our choice of leucine, a commonly used radiolabeled amino acid, may confound the decay of proteins as it has been reported that tritiated leucine is reutilized for another turn of protein synthesis (e.g. Forgue and Dahl, 1978). Altogether, what is proposed here is that the depletion in protein repertoire to a degree that de-stabilizes LTP cannot be achieved before PSI application interval is long enough beyond the half-life of a set of proteins that is necessary for LTP stabilization. Given that short time interval of PSI application has not yielded an effect on LTP but such effect was obvious with longer treatment regime, it could be assumed that the half-life of such set is longer than 30 min, in our hand. However, when short treatment duration of CHX was associated with an acceleration of protein turn-over rate resulting from peroxide-mediated protein modification, LTP became more vulnerable. This is likely to be explained by CHX effect of curtailing the replenishment processes of those degraded proteins. However, presently, this scenario can only be seen as a working hypothesis. Whether the sensitivity of LTP for short time PSI protocols, as reported in the literature, might be explained by changing the kinetic of proteins, i.e. shortening half-life, remains to be systematically investigated. Acknowledgments—I would like to thank Dr. Jo¨rgen Ekstro¨m for providing facilities for tritiated leucine incorporation assay; and Ann-Christine Reinhold for excellent technical assistance. This study was supported by the LUA/ALF-agreement (ALFGBG-
A.-K. Abbas / Neuroscience 246 (2013) 301–311 11907) and the foundations of Wilhelm & Martina Lundgren, Lennanders, and Handlanden Hjalmar Svensson; the Swedish Medical Society and the Gothenburg Medical Society.
REFERENCES Abbas A-K, Dozmorov M, Li R, Huang F-S, Hellberg F, Danielson J, Tian Y, Ekstro¨m J, Sandberg M, Wigstro¨m H (2009) Persistent LTP without triggered protein synthesis. Neurosci Res 63:59–65. Abbas A-K, Huang F-S, Li R, Ekstro¨m J, Wigstro¨m H (2011) Emetine treatment masks initial LTP without affecting long-term stability. Brain Res 1426:18–29. Abraham WC, Williams JM (2008) LTP maintenance and its protein synthesis-dependence. Neurobiol Learn Mem 89:260–268. Dozmorov M, Niu Y-P, Xu H-P, Xiao M-Y, Li R, Sandberg M, Wigstro¨m H (2003) Active decay of composite excitatory postsynaptic potentials in hippocampal slices from young rats. Brain Res 973:44–55. Engert F, Bonhoeffer T (1999) Dendritic spine changes associated with hippocampal long-term synaptic plasticity. Nature 399:66–70. Fifkova E, Anderson CL, Young SJ, VanHarreveld A (1982) Effect of anisomycin of stimulation-induced changes in dendritic spines of the dentate granule cells. J Neurocytol 11:183–210. Fligiel SEG, Lee EC, McCoy JP, Johnson KJ, Varani J (1984) Protein degradation following treatment with hydrogen peroxide. Am J Pathol 115:418–425. Fonseca R, Na¨gerl UV, Morris RGM, Bonhoeffer T (2004) Competing for memory: hippocampal LTP under regimes of reduced protein synthesis. Neuron 44:1011–1020. Fonseca R, Vabulas RM, Hartl FU, Bonhoeffer T, Na¨gerl UV (2006) A balance of protein synthesis and proteasome-dependent degradation determines the maintenance of LTP. Neuron 52:239–245. Forgue ST, Dahl JL (1978) The turnover rate of tubulin in rat brain. J Neurochem 31:1289–1297. Frey U, Morris RGM (1997) Synaptic tagging and long-term potentiation. Nature 385:533–536. Frey U, Morris RGM (1998) Weak before strong: dissociating synaptic tagging and plasticity-factor accounts of late-LTP. Neuropharmacology 37:545–552. Gieseg SP, Simpson JA, Charlton TS, Duncan MW, Dean RT (1993) Protein-bound 3,4-dihydroxyphenylalanine is a major reductant formed during hydroxyl radical damage to proteins. Biochemistry 32:4780–4786. Grune T, Reinheckel T, Joshi M, Davies JA (1995) Proteolysis in cultured liver epithelial cells during oxidative stress. Role of the multicatalytic proteinase complex, proteasome. J Biol Chem 270:2344–2351.
311
Grune T, Reinheckel T, Davies JA (1996) Degradation of oxidized proteins in K562 human hematopoietic cells by proteasome. J Biol Chem 271:15504–15509. Hesse GW, Hofstein R, Shashoua VE (1984) Protein release from hippocampus in vitro. Brain Res 305:61–66. Hunt JV, Simpson JA, Dean RT (1988) Hydroperoxide-mediated fragmentation of proteins. Biochem J 250:87–93. Larson J, Wong D, Lynch G (1986) Patterned stimulation at the theta frequency is optimal for the induction of hippocampal long-term potentiation. Brain Res 368:347–350. Lipton P, Heimbach CJ (1978) Mechanisms of extracellular potassium stimulation of protein synthesis in guinea pig hippocampal slices. J Neurochem 31:1299–1307. Lynch MA (2004) Long-term potentiation and memory. Physiol Rev 84:87–136. Maletic-Savatic M, Malinow R, Svoboda K (1999) Rapid dendritic morphogenesis in CA1 hippocampal dendrites induced by synaptic activity. Science 283:1923–1927. Martin SJ, Grimwood PD, Morris RGM (2000) Synaptic plasticity and memory: an evaluation of the hypothesis. Annu Rev Neurosci 23:649–711. Matsuzaki M, Honkura N, Ellis-Davies GCR, Kasai H (2004) Structural basis of long-term potentiation in single dendritic spines. Nature 429:761–766. Ris L, Villers A, Godaux E (2009) Synaptic capture-mediated longlasting long-term potentiation is strongly dependent on mRNA translation. NeuroReport 20:1572–1576. Sajikumar S, Frey JU (2004) Late-associativity, synaptic tagging, and the role of dopamine during LTP and LTD. Neurobiol Learn Mem 82:12–25. Shetty PK, Huang FL, Huang K-P (2008) Ischemia-elicited oxidative modulation of Ca2+/calmodulin-dependent protein kinase II. J Biol Chem 283:5389–5401. Toni N, Buchs P-A, Nikonenko I, Bron CR, Muller D (1999) LTP promotes formation of multiple spine synapses between a single axon terminal and a dendrite. Nature 402:421–425. Tsuriel S, Geva R, Zamorano P, Dresbach T, Boeckers T, Gundelfinger ED, Garner CC, Ziv NE (2006) Local sharing as a predominant determinant of synaptic matrix molecular dynamics. PLoS Biol 4:1572–1587. Villers A, Godaux E, Ris L (2010) Late phase of L-LTP elicited in isolated CA1 dendrites cannot be transferred by synaptic capture. NeuroReport 21:210–215. Villers A, Godaux E, Ris L (2012) Long-lasting LTP requires neither repeated trains for its induction nor protein synthesis for its development. PLoS One 7:e40823. Volianskis A, Jensen MS (2003) Transient and sustained types of long-term potentiation in the CA1 area of the rat hippocampus. J Physiol 550:459–492.
(Accepted 2 May 2013) (Available online 16 May 2013)