Differential presynaptic modulation of excitatory and inhibitory autaptic currents in cultured hippocampal neurons

Brain Research 1012 (2004) 22 – 28 www.elsevier.com/locate/brainres

Research report

Differential presynaptic modulation of excitatory and inhibitory autaptic currents in cultured hippocampal neurons Liu Lin Thio, Kelvin A. Yamada * Department of Neurology and the Center for the Study of Nervous System Injury, Washington University School of Medicine, 660 South Euclid Avenue, Box 8111, St. Louis, MO 63110, USA Department of Pediatrics, Washington University School of Medicine, 660 South Euclid Avenue, Box 8111, St. Louis, MO 63110, USA Division of Pediatric Neurology and the Pediatric Epilepsy Center, St. Louis Children’s Hospital, St. Louis, MO 63110, USA Accepted 28 February 2004 Available online 10 May 2004

Abstract Short-term synaptic plasticity has an important role in higher cortical function. Hyperpolarization may effect a form of short-term plasticity by promoting recovery from sodium channel inactivation or by activating axonal A-type potassium channels. To determine whether one or both processes occur, we examined the effect of hyperpolarizing prepulses on autaptic currents in cultured postnatal rat hippocampal neurons. As expected of enhanced recovery from sodium channel inactivation, hyperpolarizing prepulses reversibly increased fast excitatory autaptic currents (eacs) mediated by a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPARs), slow eacs mediated by Nmethyl-D-aspartate receptors (NMDARs), and inhibitory autaptic currents (iacs) mediated by g-aminobutyric acidA receptors (GABAARs). Hyperpolarizing prepulses augmented nearly all fast and slow eacs but only half of the iacs. This change occurred without a change in autaptic current kinetics. Of note, hyperpolarizing prepulses did not significantly reduce autaptic currents in any neuron studied. The rapidly dissociating competitive antagonists kynurenate and L-2-amino-5-phosphonovaleric acid (LAPV) inhibited fast and slow eacs, respectively, to the same extent with and without a hyperpolarizing prepulse. In addition, hyperpolarizing prepulses revealed a slow eac even after the slow eac evoked without a prepulse was completely blocked by the open channel blocker, (+)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine hydrogen maleate (MK-801). Finally, hyperpolarizing prepulses did not alter currents evoked by exogenous applications of glutamate and GABA. These findings suggest that hyperpolarizing prepulses preferentially enhance eacs over iacs, and that they do so, in part, by overcoming conduction block or by activating silent synapses. D 2004 Elsevier B.V. All rights reserved. Theme: Excitable membranes and synaptic transmission Topic: Presynaptic mechanisms Keywords: Short-term plasticity; Spillover; Silent synapse; Conduction block; Hyperpolarizing prepulse

1. Introduction Higher cortical function, such as learning and memory, partly reflects two forms of activity dependent synaptic plasticity at chemical synapses [12]. Short-term synaptic plasticity refers to changes in synaptic efficacy lasting for tens of milliseconds to minutes while long-term plasticity occurs on a time scale of hours or longer. Short-term * Corresponding author. Department of Neurology, Washington University School of Medicine, 660 South Euclid, Box 8111, St. Louis, MO 63110,USA. Tel.: +1-314-362-3585; fax: +1-314-362-9462. E-mail address: [email protected] (K.A. Yamada). 0006-8993/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2004.02.077

plasticity includes facilitation, depression, augmentation, and post-tetanic potentiation [31,33]. These processes appear to be important to both the normal function of neuronal circuits [1,5,11 – 13,28] and in promoting and inhibiting pathological processes such as seizures [16,28]. Conduction block or the failure of an action potential to propagate down all branches of the action potential is a form of short-term plasticity that occurs in invertebrates [3] and in the mammalian spinal cord [32]. Although several studies indicate that it does not occur in mammalian cortical [6,18,19] and hippocampal neurons [2,14,18,25,26], others indicate that conduction block can contribute to hippocampal synaptic failure [7,17,22] and depression [4]. If axonal

L.L. Thio, K.A. Yamada / Brain Research 1012 (2004) 22–28

conduction block occurs, it may be overcome by hyperpolarizing prepulses that enhance sodium channel recovery from inactivation [14]. The resulting action potential may then propagate down more axonal branches and produce a larger synaptic response. However, hyperpolarizing prepulses in some CA3 hippocampal neurons may promote conduction block by activating an axonal A-type potassium conductance [7,17] and produce a smaller synaptic response. To determine if one of these two processes predominates in both excitatory and inhibitory neurons, we examined the effect of hyperpolarizing prepulses on autaptic currents in cultured postnatal rat hippocampal neurons. We found that hyperpolarizing prepulses enhanced both excitatory (eacs) and inhibitory autaptic (iacs) currents but that eac enhancement occurred more consistently. Hyperpolarizing prepulses did not reduce either eacs or iacs. A preliminary version of some of these findings has been published as an abstract [29].

2. Materials and methods 2.1. Postnatal rat hippocampal cultures Microcultures and mass cultures of postnatal rat hippocampal neurons were prepared as described previously [15]. Neurons were cultured for 5 –20 days. 2.2. Whole-cell patch clamp recordings Autaptic currents were studied in microcultures containing single neurons with no neurites extending beyond the glial microisland. Currents evoked by exogenous applications of glutamate and g-aminobutyric acid (GABA) were examined in mass cultures. Whole-cell currents were recorded with an Axopatch 200A amplifier (Axon Instruments, Union City, CA). Autaptic currents were elicited by subjecting neurons voltage clamped at 60 mV to a 1-ms depolarizing voltage step to 0 mV. A holding potential of 60 mV was used in all experiments because the neurons had a resting membrane potential of 57 F 2 mV (mean F standard error of the mean) (n = 10). Series resistance compensation was set at 90 –95%, and the four-pole low pass filter on the amplifier was set at 1 – 5 kHz. All experiments were performed at room temperature. When recording autaptic currents, the neurons were bathed in an extracellular solution containing (in mM) 140 NaCl, 5 KCl, 0.75 or 2 CaCl2, 1 MgCl2, 10 D-glucose, and 10 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES) (pH 7.36). Pipettes had resistances of 3– 6 MV when filled with a solution containing (in mM) 140 KCl, 4 NaCl, 0.5 CaCl2, 5 ethylene glycol-bis(2-aminoethylether)N,N,NV,NV-tetraacetic acid (EGTA), 2 MgATP, 0.5 NaGTP, and 10 HEPES (pH 7.36). On some occasions, the KCl was replaced with K acetate or K gluconate. To examine currents elicited by 300-ms applications of glutamate or GABA, the above solutions were modified

23

as follows. The extracellular solution contained 0.5 AM tetrodotoxin and in some experiments also contained 10 mM tetraethylammonium chloride. CsCl was substituted for KCl in the internal solution while MgATP and NaGTP were omitted. Neurons were perfused continuously with either the extracellular solution or a test solution using a flow tube system. Drugs were dissolved in the extracellular solution. The bicuculline solution contained 0.1% dimethyl sulfoxide (DMSO), which had no effect on iacs. Iacs in the presence of 0.1% DMSO were 112 F 10% (n = 3) of control. 2.3. Data analysis Currents were digitized at 5 –10 kHz and analyzed using pCLAMP (Axon Instruments). Unless otherwise indicated, the experiments were performed by interleaving three to five consecutive autaptic currents under control conditions with three to five consecutive autaptic currents under experimental conditions. Usually autaptic currents were evoked at 0.05 Hz (range 0.03 – 0.1 Hz). Higher stimulation frequencies reversibly decreased autaptic currents. In most cases, neurons were subjected to the experimental condition three to five times. The data were analyzed by averaging the three to five consecutive traces obtained under either control or experimental conditions. Averaged traces are shown in the figures. For 95% or more of the cells included in this study, each stimulus produced a clear autaptic current. The decay of the averaged autaptic current was fitted to the sum of one or more exponential functions using a Simplex method. The fit was started when the current had decayed by 5%. The number of exponentials was determined by using an F-test ( p < 0.05) and by examining residual plots. The amount of charge transferred by the averaged autaptic currents was determined by calculating the area under the curve. The time for the autaptic current to rise from 20% to 80% of the peak (trise) was measured in those neurons in which the onset of the current showed little contamination by the tail current of the depolarizing step. In these cells, the time for the current to decay from 80% to 20% of the peak (tdecay) was also measured. Statistical analysis was performed using Origin (Microcal Software, Northampton, MA). Data are presented as the mean F stanstandard error of the mean. 2.4. Materials All chemicals were obtained from Sigma (St. Louis, MO) except (+)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine hydrogen maleate (MK-801), which was obtained from RBI (Natick, MA), and 1,2,3,4-tetrahydro-6-nitro-2,3-dioxo-benzo[f]quinoxaline-7sulfonamide (NBQX), which was generously provided by Lars Nordholm at Novo Nordisk (Denmark).

24

L.L. Thio, K.A. Yamada / Brain Research 1012 (2004) 22–28

3. Results 3.1. Kinetics and pharmacology of autaptic currents We first characterized the kinetic and pharmacological properties of the autaptic currents in our postnatal rat hippocampal microcultures. Fast eacs (see Fig. 1A) mediated by a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic

acid receptors (AMPARs) were sensitive to NBQX and insensitive to bicuculline. In a Mg2 +-containing but nominally glycine-free extracellular solution, 10 AM NBQX + 100 AM D,L-2-amino-5-phosphonovaleric acid (APV) reversibly decreased fast eac peak amplitudes by 99 F 1% (n = 4). Under these conditions, 10 AM NBQX alone reversibly reduced fast eac peak amplitudes by 98 F 1% (n = 5). Thus, N-methyl-D-aspartate receptors (NMDARs) did not contribute significantly to the eacs recorded under these conditions. Fast eacs were 102 F 8% (n = 6) of control in the presence of 10 AM bicuculline. They decayed by either a one or two exponential time course. The single or faster component had a time constant of 3.6 F 0.4 ms (n = 9) while the slower component had a time constant of 17 F 6.1 ms (n = 3) and accounted for 14 F 12% (n = 3) of the amplitude. These decays were similar to those reported previously [8]. g-Amino butyric acidA receptor (GABAAR)-mediated iacs (see Fig. 1A) were sensitive to bicuculline and insensitive to NBQX + APV. Ten micromolar bicuculline reversibly decreased iac peak amplitudes by 95 F 3% (n = 10). Peak amplitudes were 99 F 5% (n = 4) of control in the presence of 10 AM NBQX + 100 AM D,L-APV. Iacs decayed by either a two or three exponential time course. The predominant component had a time constant of 22 F3.4 ms (n = 9) and contributed 59 F 9% (n = 9) of the amplitude. The other two time constants were 2.7 F 0.6 ms (n = 7) and 102 F 23 ms (n = 8) accounting for 9 F 2% (n = 7) and 38 F 9% (n = 8) of the amplitude, respectively. Similar decay rates have been reported previously [15]. The difference in eac and iac decay kinetics allowed us to distinguish them reliably by visual inspection. 3.2. Hyperpolarizing prepulses reversibly augmented autaptic currents

Fig. 1. Hyperpolarizing prepulses reversibly increased fast excitatory autaptic currents (eacs) mediated by AMPARs, slow eacs mediated by NMDARs, and inhibitory autaptic currents (iacs) mediated by GABAARs. (A) Autaptic currents were elicited by 1-ms depolarizing voltage steps from 60 to 0 mV before (left column), with (middle column), and after (right column) a 20 – 80-ms hyperpolarizing prepulse to – 90 mV (top row). The traces in the right column were obtained starting 20 s after those in the middle column. Fast eacs (second row), slow eacs (third row), and iacs (bottom row) were elicited from three different neurons. NMDAR-mediated autaptic currents were isolated by using an extracellular solution containing 10 AM NBQX, 10 AM bicuculline, 5 AM glycine, and no added Mg2 +. The external [CaCl2] was 2 mM for the fast eacs and 0.75 mM for the slow eacs and iacs. (B) Percent increase in peak amplitude (solid bars) and total charge transferred (open bars) for fast eacs (left bars, n = 22 for peak and n = 20 for charge transfer), slow eacs (middle bars, n = 19 for peak and n = 17 for charge transfer), and iacs (right bars, n = 12) in cells in which hyperpolarizing prepulses increased the autaptic current. The transients in all figures were truncated for display.

We examined the effect of 20 –80-ms hyperpolarizing prepulses to 90 mV on autaptic currents. In all neurons tested with fast eacs, such prepulses increased fast eac amplitudes and the amount of charge transferred (Fig. 1A and B). NMDAR-mediated slow eacs were isolated by adding 10 AM NBQX, 10 AM bicuculline, and 5 AM glycine to the extracellular solution while excluding Mg2 +. In 94% (19/20) of neurons with a slow eac, a hyperpolarizing prepulse increased the autaptic current amplitude and the amount of charge transferred (Fig. 1A and B). In one neuron, a hyperpolarizing prepulse decreased the peak slow eac amplitude by 15% and the charge transferred by 17%. In contrast, hyperpolarizing prepulses had a similar effect in only 50% (12/24) of neurons with an iac (Fig. 1A and B). In the remaining neurons examined with an iac (12/24), the prepulse had no significant effect producing a 6 F 3% (n = 12) and 1 F 3% (n = 12) change in amplitude and the amount of charge transferred, respectively. Hyperpolarizing prepulses did not alter the trise or tdecay times of fast eacs, slow eacs,

L.L. Thio, K.A. Yamada / Brain Research 1012 (2004) 22–28

25

or iacs when they were enhanced by the prepulses (Fig. 2). Although we did not study the time course of recovery in detail, the augmenting effect of the hyperpolarizing prepulses reversed within 20 s (Fig. 1A, right panels). We performed the following experiments to determine if hyperpolarizing prepulses augment autaptic currents through a pre- or postsynaptic mechanism. 3.3. Hyperpolarizing prepulses did not change the degree of block of autaptic currents by rapidly dissociating antagonists Hyperpolarizing prepulses may enhance autaptic currents by increasing the amount of neurotransmitter released at an autapse and in turn increasing the number of receptors activated. If the hyperpolarizing prepulses increase the amount of neurotransmitter released, then a rapidly dissociating antagonist should inhibit autaptic currents elicited without a prepulse more than those elicited with a prepulse [9,30]. Two hundred micromolar kynurenate, a rapidly dissociating AMPAR antagonist, inhibited fast eacs evoked with and without a hyperpolarizing prepulse equally (Fig. 3A and C). Similarly, 100 AM L-APV, a rapidly dissociating

Fig. 3. Hyperpolarizing prepulses did not alter the degree of inhibition of fast and slow eacs by rapidly dissociating antagonists. (A) Fast eacs obtained from a single neuron without (left panel) and with (right panel) a 40-ms hyperpolarizing prepulse to 90 mV in the absence (larger currents) and presence (smaller currents marked by arrows) of 200 AM kynurenate. (B) Slow eacs obtained from a single neuron without (left panel) and with (right panel) a 40-ms hyperpolarizing prepulse to 90 mV in the absence (larger currents) and presence (smaller currents marked by arrows) of 100 AM L-APV. Voltage step protocols are shown above the current traces in A and B. (C) Inhibition of fast eacs by 200 AM kynurenate (KYNA) (n = 5) and slow eacs by 100 AM L-APV (n = 4) without (solid bars) and with (open bars) a 40-ms hyperpolarizing prepulse to 90 mV. The external [CaCl2] was 0.75 mM.

NMDAR antagonist, inhibited slow eacs elicited with and without a hyperpolarizing prepulse equally (Fig. 3B and C). These results suggest that hyperpolarizing prepulses do not augment autaptic currents by increasing the amount of neurotransmitter released at those autapses activated without a prepulse. Fig. 2. Hyperpolarizing prepulses did not alter autaptic current kinetics. The trise (left column) and tdecay (right column) for fast eacs (top row, n = 6), slow eacs (middle row, n = 4 – 5), and iacs (bottom row, n = 6) without (filled bars) and with (open bars) a hyperpolarizing prepulse. The striped bars in the middle row show the times for slow eacs (n = 3) elicited after MK-801 block as shown in Fig. 4D. trise is the time for the autaptic current to rise from 20% to 80% of the peak. tdecay is the time for the current to decay from 80% to 20% of the peak. Data were obtained from neurons in which autaptic currents were augmented by hyperpolarizing prepulses.

3.4. Hyperpolarizing prepulses revealed a slow eac after MK-801 block Another presynaptic mechanism through which hyperpolarizing prepulses may augment autaptic currents is by recruiting a second set of receptors. One set is the same as those activated without a hyperpolarizing prepulse. The

26

L.L. Thio, K.A. Yamada / Brain Research 1012 (2004) 22–28

polarizing prepulse, then no autaptic response would remain. In neurons exhibiting a slow eac potentiated by a hyperpolarizing prepulse (Fig. 4A and B), we elicited slow eacs without a hyperpolarizing prepulse in the presence of 20 AM MK-801 until no slow eac occurred (Fig. 4C). Slow eacs elicited in the presence of 10 –40 AM MK-801 progressively declined in amplitude with successive stimuli until no response remained as described previously [23]. Generally, no slow eac could be seen after five stimuli. Once the slow eac was blocked, we examined the effect of a 40-ms hyperpolarizing prepulse on the slow eac recorded in the absence of MK-801. In six of eight neurons, a slow eac was consistently observed with a hyperpolarizing prepulse after the slow eac evoked without a prepulse was blocked by MK-801 (Fig. 4D). The trise and tdecay for the slow eacs evoked by a hyperpolarizing prepulse after MK-801 were the same as those for the slow eacs evoked without a hyperpolarizing prepulse (Fig. 2, middle row). Thus, the findings support the hypothesis that hyperpolarizing prepulses augment slow eacs by activating a second set of

Fig. 4. Hyperpolarizing prepulses revealed a slow eac after MK-801 completely eliminated the slow eac elicited without a hyperpolarizing prepulse. All traces were obtained from a single neuron. (A, B) Two almost identical slow eacs evoked without a hyperpolarizing prepulse are superimposed. One was obtained before (A, solid trace) and the other after (A, gray dotted trace) a slow eac evoked with a 40-ms hyperpolarizing prepulse to-go mV (B). (C) Autaptic current remaining after a 3-min exposure to 20 AM MK-801 during which autaptic currents were elicited without a hyperpolarizing prepulse at 0.05 Hz. (D) After completely blocking the autaptic current with MK-801, the hyperpolarizing prepulse paradigm elicited an autaptic current. Slow eacs were isolated as in Fig. 2. The external [CaCl2] was 0.75 mM.

second set may not be activated without a prepulse because of conduction block or because it is located at a silent autapse. According to Malenka and Nicoll [20], a silent autapse is structurally normal but does not function because the neurotransmitter is not released when an action potential enters the presynaptic terminal or because the postsynaptic membrane is unresponsive to the released neurotransmitter. We used the NMDA open channel blocker MK-801 to begin to examine these possibilities. We reasoned that all NMDA channels capable of contributing to a slow eac generated without a hyperpolarizing prepulse could be blocked by repeatedly evoking a slow eac in the presence of MK-801. Thus, if hyperpolarizing prepulses result in the activation of a second set of channels, then a slow eac activated with a hyperpolarizing prepulse would remain even after MK-801 completely blocked the slow eac evoked with no prepulse. Alternatively, if one set of receptors mediates the response with and without a hyper-

Fig. 5. Hyperpolarizing prepulses did not increase currents activated by exogenous applications of glutamate and GABA. The AMPAR-mediated current evoked by 200 AM glutamate (A), the NMDAR-mediated current evoked by 2 AM glutamate in an extracellular solution containing 5 AM glycine and no added Mg2 + (B), and the current evoked by 20 AM GABA (C) without (left panels) and with (right panels) a 40-ms hyperpolarizing prepulse to 90 mV. Traces were recorded from three different neurons. Neurotransmitters were applied for 300 ms. The flow tube system caused the transients seen in some traces. These transients and the capacitive transients were truncated for display. The extracellular [CaCl2] was 2 mM in A and B while it was 0.75 mM in C.

L.L. Thio, K.A. Yamada / Brain Research 1012 (2004) 22–28

receptors. However, the amplitude of and the amount of charge transferred by the remaining slow eacs were only 64 F 17% and 55 F 15%, respectively, of the expected values. A potential explanation for this finding is that some rundown occurred in these experiments. 3.5. Hyperpolarizing prepulses did not increase AMPA, NMDA, and GABA receptor-mediated responses evoked by exogenous applications of glutamate and GABA To determine whether hyperpolarizing prepulses affected postsynaptic receptors directly, we examined the effect of hyperpolarizing prepulses on currents evoked by exogenous applications of glutamate and GABA. Glutamate and GABA concentrations near the EC50 were selected [15,24]. AMPARs were assessed by examining the rapidly desensitizing current elicited by 200 AM glutamate. NMDARs were assessed by studying the current activated by 2 AM glutamate with 5 AM glycine and no added Mg2 + in the extracellular solution. The amplitudes of the AMPAR and NMDAR mediated currents after a 40-ms hyperpolarizing prepulse to 90 mV were 100 F 4% (n = 3) and 102 F 1% (n = 3) of control, respectively (Fig. 5A and B). The amplitudes of currents generated by 20 AM GABA after a 40-ms hyperpolarizing prepulse to 90 mV were 105 F 4% (n = 4) of control (Fig. 5C). These data indicate that hyperpolarizing prepulses do not directly modulate AMPARs, NMDARs, and GABAARs.

4. Discussion The primary finding of this study is that hyperpolarizing prepulses augment fast AMPAR-mediated eacs, slow NMDAR-mediated eacs, and GABAAR-mediated iacs. Hyperpolarizing prepulses may augment autaptic currents either by activating additional receptors, by altering the properties of the activated receptors, or a combination of both mechanisms. The prepulses elicited a slow eac even after MK-801 had completely blocked the slow eac elicited in the absence of a hyperpolarizing prepulse. This result suggests that the autaptic currents evoked with a prepulse reflect the activation of two sets of receptors— one identical to those activated without a prepulse and a second at another location. The second receptor set only contributes to the autaptic current elicited with a prepulse because without a prepulse, a sufficient concentration of the released transmitter does not reach these receptors. Alternatively, the released transmitter binds to these receptors but fails to activate them in the absence of a hyperpolarizing prepulse. According to this model, the hyperpolarizing prepulses must act postsynaptically to alter AMPARs, NMDARs, and GABAARs. Although we found no evidence for such a mechanism (Fig. 5), we cannot exclude the possibility that synaptic and extrasy-

27

naptic receptors respond differently to hyperpolarizing prepulses. The additional receptors contributing to autaptic currents evoked with a hyperpolarizing prepulse may be extrasynaptic. One mechanism for opening extrasynaptic receptors is spillover caused by increased transmitter release. Hyperpolarizing prepulses may increase transmitter release by enhancing any of the steps involved in calcium-dependent transmitter release from presynaptic terminals. In this case, the extrasynaptic transmitter concentration must be lower than the transmitter concentration in the synaptic cleft. Thus, autaptic currents activated with a hyperpolarizing prepulse should be inhibited to a greater extent by the rapidly dissociating antagonists kynurenate and L-APV than those activated without a prepulse [10]. Second, spillover causing the activation of extrasynaptic receptors should slow the activation and decay rates of the autaptic currents [10,27]. Indeed, the kinetics of the slow eacs evoked with a hyperpolarizing prepulse after MK-801 block were the same as those evoked without a prepulse. Since neither prediction is true experimentally (Figs. 2 and 3), we believe that spillover does not account for our results. A second possible location for the additional receptors activated with a hyperpolarizing prepulse are autapses that for one of two reasons are not otherwise activated. The prepulses may activate presynaptically silent terminals by enhancing calcium entry or any of the processes involved in coupling calcium entry to vesicular release. Alternatively, hyperpolarizing prepulses may activate additional autapses by overcoming axonal conduction block, which others have concluded occurs in the hippocampus [4,7,17,22]. Hyperpolarizing prepulses may overcome conduction block by promoting recovery of sodium channels from the inactivated state. Our results are consistent with either mechanism. However, our results are not compatible with the activation of an axonal A-type potassium conductance by the hyperpolarizing prepulses as Debanne et al. [7] found in CA3 neurons in hippocampal slice cultures. If hyperpolarizing prepulses activated this conductance in our preparation, they should have decreased autaptic currents in at least some neurons. We may not have observed any axonal A-type potassium current effects because our studies were performed before cultured hippocampal neurons express one of the axonal A-type potassium channel polypeptides [21]. In summary, we found that hyperpolarizing prepulses reveal a form of short-term plasticity that enhances eacs more effectively than iacs via a presynaptic mechanism involving the activation of silent synapses or overcoming conduction block. Thus, this form of short-term plasticity shows a preference for glutamatergic neurons. Precedence for selective presynaptic modulation of excitatory synapses comes from the observation that sodium channel blockers depress eacs to a greater degree than iacs [25] and that elevated potassium affects paired-pulse modulation of eacs but not iacs [14,22]. Together, these results indicate that axonal action potential conduction, the coupling of action

28

L.L. Thio, K.A. Yamada / Brain Research 1012 (2004) 22–28

potentials to vesicular release, or both differ between glutamatergic and GABAergic neurons.

Acknowledgements National Institutes of Health Neurological Sciences Academic Development Award 5K12NS0169004 supported this work. We thank Matthew W. Hill, Steven Mennerick, Steven M. Rothman, and Charles F. Zorumski for helpful comments and discussion. We thank Nancy Lancaster for preparing and maintaining the neuronal cultures.

[15]

[16]

[17]

[18]

[19]

References [1] L.F. Abbott, J.A. Varela, K. Sen, S.B. Nelson, Synaptic depression and cortical gain control, Science 275 (1997) 220 – 224. [2] C. Allen, C.F. Stevens, An evaluation of causes for unreliability of synaptic transmission, Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 10380 – 10383. [3] S.A. Baccus, B.D. Burrell, C.L. Sahley, K.J. Muller, Action potential reflection and failure at axon branch points cause stepwise changes in EPSPs in a neuron essential for learning, J. Neurophysiol. 83 (2000) 1693 – 1700. [4] D.L. Brody, D.T. Yue, Release-independent short-term synaptic depression in cultured hippocampal neurons, J. Neurosci. 20 (2000) 2480 – 2494. [5] D.V. Buonomano, M.M. Merzenich, Temporal information transformed into a spatial code by a neural network with realistic properties, Science 267 (1995) 1028 – 1030. [6] C.L. Cox, W. Denk, D.W. Tank, K. Svoboda, Action potentials reliably invade axonal arbors of rat neocortical neurons, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 9724 – 9728. [7] D. Debanne, N.C. Gue´rineau, B.H. Ga¨hwiler, S.M. Thompson, Actionpotential propagation gated by an axonal IA-like K+ conductance in hippocampus [published erratum appears in Nature 1997 Dec 4;390(6659):536], Nature 389 (1997) 286 – 289. [8] J.S. Diamond, C.E. Jahr, Asynchronous release of synaptic vesicles determines the time course of the AMPA receptor-mediated EPSC, Neuron 15 (1995) 1097 – 1107. [9] J.S. Diamond, C.E. Jahr, Transporters buffer synaptically released glutamate on a submillisecond time scale, J. Neurosci. 17 (1997) 4672 – 4687. [10] D.A. DiGregorio, Z. Nusser, R.A. Silver, Spillover of glutamate onto synaptic AMPA receptors enhances fast transmission at a cerebellar synapse, Neuron 35 (2002) 521 – 533. [11] L.E. Dobrunz, C.F. Stevens, Response of hippocampal synapses to natural stimulation patterns, Neuron 22 (1999) 157 – 166. [12] S.A. Fisher, T.M. Fischer, T.J. Carew, Multiple overlapping processes underlying short-term synaptic enhancement, Trends Neurosci. 20 (1997) 170 – 177. [13] E.S. Fortune, G.J. Rose, Short-term synaptic plasticity as a temporal filter, Trends Neurosci. 24 (2001) 381 – 385. [14] Y. He, C.F. Zorumski, S. Mennerick, Contribution of presynaptic Na+

[20] [21]

[22]

[23]

[24]

[25]

[26] [27]

[28] [29]

[30] [31] [32] [33]

channel inactivation to paired-pulse synaptic depression in cultured hippocampal neurons, J. Neurophysiol. 87 (2002) 925 – 936. M.W. Hill, P.A. Reddy, D.F. Covey, S.M. Rothman, Contribution of subsaturating GABA concentrations to IPSCs in cultured hippocampal neurons, J. Neurosci. 18 (1998) 5103 – 5111. M. Kokaia, F. Asztely, K. Olofsdotter, C.B. Sindreu, D.M. Kullmann, O. Lindvall, Endogenous neurotrophin-3 regulates short-term plasticity at lateral perforant path-granule cell synapses, J. Neurosci. 18 (1998) 8730 – 8739. I.L. Kopysova, D. Debanne, Critical role of axonal A-type K+ channels and axonal geometry in the gating of action potential propagation along CA3 pyramidal cell axons: a simulation study, J. Neurosci. 18 (1998) 7436 – 7451. P.J. Mackenzie, T.H. Murphy, High safety factor for action potential conduction along axons but not dendrites of cultured hippocampal and cortical neurons, J. Neurophysiol. 80 (1998) 2089 – 2101. P.J. Mackenzie, M. Umemiya, T.H. Murphy, Ca2 + imaging of CNS axons in culture indicates reliable coupling between single action potentials and distal functional release sites, Neuron 16 (1996) 783 – 795. R.C. Malenka, R.A. Nicoll, Silent synapses speak up, Neuron 19 (1997) 473 – 476. M. Maletic-Savatic, N.J. Lenn, J.S. Trimmer, Differential spatiotemporal expression of K+ channel polypeptides in rat hippocampal neurons developing in situ and in vitro, J. Neurosci. 15 (1995) 3840 – 3851. J.P. Meeks, S. Mennerick, Selective effects of potassium elevations on glutamate signaling and action potential conduction in hippocampus, J. Neurosci. 24 (2004) 197 – 206. V.N. Murthy, T.J. Sejnowski, C.F. Stevens, Heterogeneous release properties of visualized individual hippocampal synapses, Neuron 18 (1997) 599 – 612. D.K. Patneau, M.L. Mayer, Structure – activity relationships for amino acid transmitter candidates acting at N-methyl-D-aspartate and quisqualate receptors, J. Neurosci. 10 (1990) 2385 – 2399. M. Prakriya, S. Mennerick, Selective depression of low-release probability excitatory synapses by sodium channel blockers, Neuron 26 (2000) 671 – 682. M. Raastad, G.M. Shepherd, Single-axon action potentials in the rat hippocampal cortex, J. Physiol. 548 (2003) 745 – 752. D.J. Rossi, M. Hamann, Spillover-mediated transmission at inhibitory synapses promoted by high affinity alpha6 subunit GABAA receptors and glomerular geometry [published erratum appears in Neuron 1998 Jul;21(1):527], Neuron 20 (1998) 783 – 795. M. Steriade, Coherent oscillations and short-term plasticity in corticothalamic networks, Trends Neurosci. 22 (1999) 337 – 345. L.L. Thio, K.A. Yamada, Branch point failure causes silent autapses in cultured postnatal rat hippocampal neurons, Abstr.-Soc. Neurosci. 24 (1998) 566. G. Tong, C.E. Jahr, Multivesicular release from excitatory synapses of cultured hippocampal neurons, Neuron 12 (1994) 51 – 59. H. Von Gersdorff, J.G. Borst, Short-term plasticity at the calyx of held, Nat. Rev., Neurosci. 3 (2002) 53 – 64. P.D. Wall, Do nerve impulses penetrate terminal arborizations? A prepresynaptic control mechanism, Trends Neurosci. 18 (1995) 99 – 103. R.S. Zucker, W.G. Regehr, Short-term synaptic plasticity, Annu. Rev. Physiol. 64 (2002) 355 – 405.