- Email: [email protected]
Enduring changes in Purkinje cell electrophysiology following transient exposure to AMPA: correlates to dark cell degeneration
Neuroscience Research 33 (1999) 155 – 162
Enduring changes in Purkinje cell electrophysiology following transient exposure to AMPA: correlates to dark cell degeneration Jean C. Strahlendorf a,*, Howard K. Strahlendorf b b
a Texas Tech Uni6ersity, Health Science Center, Department of Physiology, 3601 4th St., Lubbock, TX 79430, USA Texas Tech Uni6ersity, Health Science Center, Department of Pharmacology, 3601 4th St., Lubbock, TX 79430, USA
Received 3 August 1998; accepted 11 December 1998
Abstract Purkinje cells (PCs) are selectively vulnerable to a-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA)-mediated delayed toxicity that is manifested as dark cell degeneration (DCD) rather than necrosis. The purpose of the present study was to utilize electrophysiologic changes induced by AMPA to gain mechanistic insights into its cytotoxic actions. The whole-cell configuration of the patch clamp technique was used to record spontaneous electrical activity and ionic currents of Purkinje neurons from cerebellar slices using an experimental paradigm known to produce DCD in response to AMPA. Initial electrophysiologic responses to AMPA consisted of a large transient depolarization and inward current that declined by 75% 20 min into the 30-min exposure to 30 mM AMPA. Cellular responses temporarily continued towards basal levels following removal of AMPA. A sustained membrane depolarization (and underlying persistent inward current), an abundance of apparent excitatory synaptic events, and loss of electro- and chemoresponsiveness were observed 60 – 75 min into the expression phase (following AMPA removal). These events correspond temporally to the development of DCD in Purkinje cells and may represent an electrophysiological signature of AMPA receptor-mediated delayed neurotoxic events. Antagonists of the AMPA receptor present concomitantly with AMPA are known not to affect DCD and failed to alter the electrophysiologic changes. The secondary depolarization and loss of electroresponsiveness were prevented by antagonists present after removal of AMPA, at a time when DCD also is prevented. Electrical clamping of the PC membrane to equivalent depolarized membrane potentials (Vms) obtained with AMPA failed to elicit any long lasting alterations in PC physiology. Collectively, morphological and electrophysiological data indicate that induction of DCD is not strongly dependent on ionotropic mechanisms elicited by AMPA receptors, but that expression of DCD does possess an ionotropic element. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: AMPA; Glutamate; Excitatory amino acids; Purkinje cells; Neurotoxicity; Electrophysiology
1. Introduction Although glutamate (glu) is benign as an excitatory transmitter in normal concentrations, abnormal glutamatergic tone may play a role in both acute neurotoxicity and chronic neurodegenerative processes. A sudden massive glu release has been linked intimately to acute neurotoxicity associated with cerebral ischemia, * Corresponding author. Tel.: +1-806-743-2554; fax: + 1-806-7431512.; e-mail: [email protected].
seizures, hypoglycemia, and trauma (Meldrum and Garthwaite, 1990). In contrast, mechanisms of chronic neurodegenerative diseases including hereditary ataxias are less understood but may be related to slow excitotoxic processes initiated inter alia by an impairment of energy metabolism, excitatory amino acid (EAA) receptor abnormalities, and slow or periodic excessive glutamatergic tone (Beal, 1992). It is now clear that excitotoxicity is multidimensional involving multiple glu receptor subtypes and mechanisms that produce several neurotoxic profiles. A wide-
0168-0102/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 8 - 0 1 0 2 ( 9 8 ) 0 0 1 2 6 - 6
156
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
spread salient role for AMPA receptors in delayed or latent excitotoxicity that equals or exceeds the role of NMDA receptors has emerged recently. Indeed, AMPA receptor-induced degeneration most often elicits a delayed toxicity (Garthwaite and Garthwaite, 1990; Nellgard and Wieloch, 1992), and unlike the widely studied mechanisms underlying acute edematous necrosis (EN) often mediated by abusive NMDA receptor activation, cellular processes associated with delayed degeneration remain relatively obscure. Minimally, delayed neuronal death is a multistage process in which transient exposure to high concentrations of EAAs serves as an initiation, induction or trigger event that promotes maturation of secondary cytotoxic cascades (Meldrum and Garthwaite, 1990). Thus, in delayed neurotoxicity, a striking temporal decoupling of toxic exposure from death is evident. Purkinje cells (PCs) express a paucity of functional NMDA receptors but a preponderance of AMPA receptors (Llano et al., 1988, 1991; Audinat et al., 1990; Krupa and Crepel, 1990; Rosenmund et al., 1992) and they are selectively vulnerable to delayed AMPA receptor-elicited toxicity (Garthwaite and Garthwaite, 1990). Therefore, the PC represents a unique neuron to study the cellular processes underlying AMPA toxicity in relative isolation of NMDA-mediated events. Delayed AMPA-induced toxicity has been demonstrated in cerebellar and hippocampal slices (Garthwaite and Garthwaite, 1991a) where AMPA was a potent neurotoxin capable of eliciting dark cell neurodegeneration (DCD) of PCs and pyramidal neurons hours after a brief exposure. AMPA-induced delayed DCD of PCs was found to be independent of both extracellular Ca2 + and AMPA antagonists, if present concomitantly with AMPA exposure; however antagonists did prevent AMPA-induced DCD when introduced after AMPA was removed (Garthwaite and Garthwaite, 1991b). This implies that an enduring AMPA receptor activation in PCs may be responsible for DCD. Electrophysiologic recordings of a population response from cerebellar cortical neurons indicated depolarization during the expression phase following AMPA removal (Garthwaite et al., 1986). Collectively, these findings indicate ionotropism is of little importance to the induction of AMPA-induced DCD in PCs, but may play a role in its expression. Although the studies of Garthwaite and colleagues (Garthwaite et al., 1986; Garthwaite and Garthwaite, 1990; Meldrum and Garthwaite, 1990) provided seminal data regarding AMPA-induced degeneration of PCs, these studies have not addressed the neurophysiology associated with the AMPA-selective delayed, degenerative process at the single PC level, the major target neuron. Changes in electrical properties of a neuron can provide a sensitive, immediate, and continuous index of the neuron’s physiology that can reflect
processes underlying cell death. Accordingly, the purpose of the present study was to gain insight into potential mechanisms underlying DCD by studying the electrophysiology of PCs subjected to AMPA receptormediated DCD.
2. Materials and methods Postnatal 8- to 12-day-old rats (Sprague–Dawley strain, Sasco Labs) were used in these studies. Sagittal slices (400 mm thick) of the vermal region of the cerebellum were obtained with a Vibroslice (Campden Instruments). Slices were allowed to recover for 1 h at room temperature submerged in artificial cerebrospinal fluid (ACSF) containing the following (in mM): NaCl (120), KCl (2), KH2PO4 (1.2), MgSO4 (1.2), CaCl2 (2), NaHCO3 (26), and glucose (11). For electrophysiologic recordings, slices were placed in an interface recording chamber, semi-submerged in ACSF flowing under and around them (1–3 ml/min) and blanketed with humidified 95:5 (%) O2/CO2. Slices were subjected to an experimental protocol that consisted of a 30-min treatment with 30 mM AMPA (termed trigger or induction phase), followed by a 90-min period in AMPA-free ACSF (expression phase). Control slices were exposed to the same protocol but in the absence of AMPA for the first 30 min. This experimental paradigm has reliably produced DCD in the majority of PCs (Garthwaite and Garthwaite, 1991a,b; Strahlendorf et al. 1996). Antagonists were superfused 10 min prior to the application of AMPA and given for the entire protocol or only during the trigger (induction) phase. The whole-cell configuration of the patch clamp technique was used to record spontaneous electrical activity under current-clamp and ionic currents under voltageclamp. Glass tubing blanks were pulled to produce electrodes with a tip resistance of 3–8 MV. The electrode solution contained (in mM): 140 K + gluconate, 2.5 MgCl2, 10 HEPES, 0.6 EGTA, 0.06 CaCl2, 4 NaATP, and 0.4 Na-GTP, pH 7.2. On occasion, Lucifer Yellow (1 mg/ml) was included in the solution to label the recorded cell for identification after the experiment. Electrical signals were led into a Dagan 3900 amplifier and displayed on a storage oscilloscope. Signals were permanently recorded on videotape using an analog to digital converter (PCM-2 A/D, Medical Systems) connected to a video recorder (Panasonic). An IBM-compatible AT computer equipped with an Axon Instruments TL-1 interface and pClamp software (Version 5.5) was used for data acquisition and analysis. Electrical recordings from PCs were made at 33°C. Only healthy, viable PCs were used to assess AMPA-induced toxicity. Viability was assessed using the following criteria: (1) a stable resting potential of at least − 55 mV for greater than 10 min, (2) the ability to fire
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
157
Fig. 1. Whole-cell current clamp recording of an individual Purkinje cell (PC) in the cerebellar slice preparation in response to AMPA (30 mM) superfused for 30 min. (A) PC membrane response during AMPA exposure (beginning at the first downward arrowhead and ending at the second downward arrowhead). AMPA elicited a depolarization with continuous Na + and Ca2 + spikes (denoted by an unfilled circle) until the membrane depolarized to a potential of approximately − 23 mV. This was accompanied by a decrease in input resistance. A sustained depolarization (indicated by an asterisk) peaked near − 18 mV. During the latter portion of the AMPA exposure period, the membrane potential remained depolarized at approximately − 30 mV and Rin began to return towards control levels. (B) PC membrane response immediately after AMPA exposure (0 – 30 min into expression phase). The PC maintained a sustained depolarization that ranged between − 25 and − 40 mV with a decrease in resistance observed during the maximal depolarization. (C) PC membrane response 30 – 60 min into the expression phase. The membrane had returned towards basal potential and Rin before initiating an increase in electrical activity. The PC membrane continued to exhibit an increase in spontaneous excitatory events with an accompanying depolarization towards − 30 mV (denoted by a crossed diamond) with a variable change in Rin. (D) PC response 60–90 min into the expression phase.
spontaneous or current-driven action potentials, (3) a reasonably high membrane resistance (\ 200 MV), and (4) the presence of hyperpolarization-activated inward current (Ih), since this current is present in all healthy PCs. Pipette capacitance was neutralized to a minimum before rupture; after rupture, the height of the capacitive transient in response to a 10-mV test pulse was used to calculate and compensate (\75%) the series resistance. Changes in magnitudes of membrane potentials (Vms), membrane currents, slope resistances, slope conductances, and electro- and chemoresponsiveness were assessed after AMPA applications. Slope resistances and conductances were calculated from I/V or V/I curves generated by 400 ms currents (10 pA incrementally) or voltage pulses (10 mV incrementally), respectively. Electroresponsiveness was evaluated at regular intervals by the capability of PCs to generate voltagedependent active membrane responses from − 60 mV. Specifically, the ability of PCs to fire action potentials or generate inward current in response to depolarizing current or voltage steps, respectively, and the capacity of PCs to elicit Ih in response to hyperpolarizing pulses was analyzed. Maintenance of chemoresponsiveness was measured by the magnitude of membrane responses (mV or pA) elicited by local AMPA applications of 10-ml microdrops, 75 mM. Loss of functional viability was defined as the inability to fire action potentials upon depolarization and loss of ligand- and/ or voltage-gated currents. Experiments in which recordings of Vm and input resistance (Rin) were unstable, series resistance could not be adequately compensated, or a sudden loss in Vm, Rin, and electro- and chemore-
sponsiveness occurred (most likely due to rupture of the plasma membrane), were not included in the data analysis. All electrophysiological parameters were normalized to pre-AMPA conditions, and post-AMPA changes expressed as percent differences. The data were expressed as the mean percentage9 S.E.M.
3. Results Control studies without AMPA superfusion (n=3) were done to establish the long-term stability of the recording procedure. These experiments revealed excellent preservation of PC electro- and chemoresponsiveness for more than 3 h. Specifically, PCs maintained membrane potentials close to − 60 mV without loss of monitored Na + - and Ca2 + -dependent action potentials and voltage-gated currents such as Ih. Additionally, PCs maintained chemical responsiveness to periodic local microapplications of AMPA, demonstrating reproducible depolarizations throughout a 3 h recording period. Loss of whole cell recordings marked by a sudden loss of Rin and a marked depolarizing shift in Vm to near 0 mV that failed to return towards control values occasionally occurred. Data from these PCs were excluded from further analysis. In current clamp mode, AMPA (30 mM, 30 min superfusion, n=13) elicited in PCs a characteristic pattern of membrane responses (Fig. 1). In all PCs during the first 5–10 min, AMPA elicited a depolarization with continuous composite Na + and Ca2 + action potentials until the membrane reached a potential that
158
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
averaged −25 ( 92 S.E.M.) mV (Fig. 1A). Subsequently, a transient 5 mV repolarization lasting approximately 15 s preceded a sustained depolarization that peaked near −10 (94) mV. The latter depolarization was accompanied by an average 31(9 6)% decrease in Rin. During the latter 10 min of the AMPA exposure period, membrane resistance returned to near control (pre-AMPA) levels, although Vm remained depolarized to an average plateau potential of − 40 (9 4) mV. During the AMPA expression period (time during which DCD is expressed), PCs remained depolarized to an average of −38 (9 5) mV, although Rin remained within 20% of the control levels (Fig. 1B). Apparent Ca2 + action potentials and compound miniature excitatory potentials (10 mV in amplitude and 15 ms in duration) appeared 40 – 50 min into the expression period (Fig. 1C). Typically, 60 – 75 min into the expression period, the Vm showed a continuously increasing sustained depolarization towards − 20 ( 9 6) mV with a maximal decrease in Rin (average 42911%). The depolarization prevailed until terminating the experiment at 90 min (Fig. 1D). During this period, manually clamping the Vm back to control values ( −55 to −60 mV) also revealed loss of spontaneous and depolarizationevoked Ca2 + - and Na + -dependent action potentials, and disappearance of voltage-gated responses (Ih) to hyperpolarizing current pulses. Prior to the 30-min AMPA exposure period, local microapplication of AMPA (75 mM, 10 ml) onto PCs produced an average 35 ( 97) mV depolarization with an average 13 (93)% decrease in Rin followed by a 6 ( 92) mV hyperpolarization and a slight increase in Rin (average 15 9 6%). Responsiveness of PCs to local microapplications of AMPA was markedly reduced by 60 – 75 min into the expression phase and at the end of the experimental protocol (i.e. at the end of the 90-min expression phase). Specifically, AMPA elicited only a 5 ( 9 2) mV depolarization with no apparent accompanying membrane resistance change or secondary hyperpolarization with the Vm manually clamped at control values. Voltage-clamp recordings revealed a characteristic profile of membrane currents in response to AMPA. During AMPA application, PCs exhibited a marked, sustained inward current that averaged 420 (956, n= 4) pA with an accompanying increase in membrane conductance averaging 74 (9 17)% of control that returned towards control values during the latter minutes of the exposure period (Fig. 2A). During the ensuing 60 min following AMPA exposure, membrane current and conductance migrated towards control levels and spontaneous, unclamped regenerative currents began to appear at times that varied between 10 and 30 min (Fig. 2B,C). A sustained inward current (average 114 918 pA) and increase in conductance (average 48916% of control) occurred approximately 60 min into the ex-
pression phase and temporally coincided with the sustained depolarization and decreased membrane resistance observed under current-clamp conditions (Fig. 2D). In the majority of PCs studied, a flurry of inward compound miniature excitatory currents preceded or occurred coincident with the onset of the sustained inward current (Fig. 2D). Loss of electroresponsiveness 60–75 min into the expression phase was also evident by failure to elicit in PCs Ih, ICa2 + or INa + following application of appropriate voltage steps from a Vm clamped at − 60 mV. To assess the contribution of ionotropic processes in mediating AMPA-induced electrophysiological events, the competitive antagonist 6-cyano-7-nitroquinoxaline2,3-dione (CNQX) (10 mM, n= 3) was superfused during the entire protocol (with AMPA and during the expression phase). This paradigm of CNQX application
Fig. 2. Whole-cell voltage clamp recording of an individual Purkinje cell (PC) in the cerebellar slice preparation in response to AMPA (30 mM) superfused for 30 min. (A) A 30-min AMPA exposure (beginning at the diamond symbol) elicited a double-peaked inward current with a transient flurry of unclamped regenerative currents (indicated by the arrowhead) and an increase in membrane conductance. During the latter minutes of AMPA exposure, a sustained inward current of approximately 150 pA was evident. (B) PC membrane current response immediately after AMPA exposure (0 – 30 min into expression phase). The PC maintained its pre-AMPA current values that ranged between 50 and 100 pA with spontaneous unclamped regenerative currents. Occasional inward plateau currents elicited bursts of regenerative currents that spontaneously subsided (denoted by an asterisk). (C) PC membrane response 30 – 60 min into the expression phase. The inward current continued at basal levels with occasional spontaneous regenerative currents and membrane conductance also had recovered to near control, pre-AMPA, values. Amplitudes of unclamped regenerative currents typically began to diminish 50 – 60 min into the expression period. (D) PC currents 60 – 90 min into the expression phase. The PC membrane continued to exhibit an increase in spontaneous excitatory regenerative currents (indicated by an asterisk) that culminated in a prolonged inward current shift of approximately 200 pA (indicated by a crossed diamond) and increased membrane conductance. This condition persisted until termination of the recording at greater than 90 min after AMPA exposure.
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
159
Fig. 3. Response of a PC to AMPA (30 mM) superfused for 30 min in the continuous presence of CNQX 10 mM using whole-cell current clamp recording. CNQX superfusion was started 10 min prior to application of AMPA. (A) In the presence of CNQX, the PC displayed a marked depolarization that peaked near −16 mV, with an associated decrease in Rin. However, unlike the response observed with AMPA only, Rin and Vm rapidly returned to near control values accompanied with continuous Na + and Ca2 + spikes within 5 min after the onset of AMPA application. Diamond denotes the beginning of the AMPA application and the bar represents the basal Vm. (B) Response of the same PC, 35–45 min into the AMPA expression phase (following AMPA removal, CNQX still present). The PC continued to spontaneously fire action potentials with a Vm slightly depolarized at approximately −53 mV and a modest increase in Rin (approximately 25% increase). (C) The response of the PC 60 – 75 min into the expression phase. As is evident, the characteristic secondary depolarization observed during the expression phase was absent when CNQX was present. (D) The response of the PC 90 min into the expression phase. Vm was approximately −60 mV with an approximate 25% increase in Rin. The PC continued to exhibit synaptic potentials and spontaneous action potentials.
previously was reported to effectively block AMPA-induced DCD (Garthwaite and Garthwaite, 1991a). In the presence of CNQX, AMPA still exhibited a marked depolarization of PCs that peaked at an average − 27 ( 9 5) mV compared to − 10 mV in the absence of CNQX (Fig. 3A). Although the associated decrease in Rin was equivalent initially to that seen in the absence of CNQX (31[ 98] vs 31[ 96]%, respectively), Rin rapidly returned to near pre-AMPA values 5 min into the AMPA application. In the presence of CNQX, the AMPA-induced plateau potential reached an average of − 51 (9 3) mV compared to −40 (9 4) mV in the absence of CNQX. In the presence of CNQX during the ensuing 60 min following the removal of AMPA, the membrane potential of PCs and accompanying input resistance remained at essentially control levels, at most evincing a 5-mV depolarization (Fig. 3B,C). The secondary depolarization observed during the expression phase was absent with CNQX present (Fig. 3D). Moreover, CNQX preserved the ability of PCs to continuously discharge spontaneous Ca2 + and Na + action potentials, generate voltage-dependent membrane responses (Ih) and display miniature excitatory potentials. Furthermore, CNQX maintained Rin at a higher level (420 MV AMPA + CNQX vs 310 MV AMPA alone) indicating that the membrane was not excessively leaky (Fig. 3B–D). To determine the contribution of ionotropic processes during the trigger phase, competitive AMPA receptor antagonists CNQX (10 mM, n =3, IC50 = 0.3 mM) or DNQX (30 and 60 mM, n = 5, IC50 =0.5 mM) were superfused with AMPA only during the 30-min
trigger phase, and then removed. Characteristic electrophysiologic events observed during the expression phase (see above) were minimally affected by the presence of either of the two antagonists at any of the given concentrations during the trigger phase (data not shown). This was not unexpected given the similarity in IC50s and previously reported morphological experiments that demonstrated conclusively that AMPA receptor antagonists present only during the trigger phase failed to block DCD (Garthwaite and Garthwaite, 1991a). Analogous to the expression phase in AMPA alone experiments, miniature excitatory potentials appeared within 15 min and increased in frequency during the following 50–60 min until regenerative spikes appeared, superimposed on a depolarizing Vm. A sustained depolarization to an average − 28 ( 9 9) mV (with an average 36 [9 12]% decrease in Rin) peaked 75 min into the expression phase. Moreover, voltage-gated currents (INa and Ih) were also blocked totally during the latter stages of the expression period. Thus, inclusion of an ionotropic EAA antagonist coincident with AMPA, but not during the expression phase, failed to prevent the characteristic delayed and sustained excitatory events during the expression period seen following application of only AMPA. A summary of electrophysiological changes observed in PCs exposed to AMPA alone and in the presence of antagonists is provided in Table 1. To ascertain whether only a 30-min sustained depolarization of the membrane alone could be sufficient to induce secondary electrophysiologic effects similar to AMPA exposure, PCs were manually clamped to Vms equivalent to those elicited by AMPA superfusion for
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
160
Table 1 Summary of changes in membrane properties of Purkinje neurons exposed to AMPA Trigger/induction phase
C-Clamp Vm (mV) Rin (% D)
AMPA
AMPA and ANTAG (trig)a
AMPA
AMPA and ANTAG (trig)a
AMPA and ANTAG (entire)b
−109 4 ¡31 96
−279 5 ¡3198
−209 6 ¡42 9 11
−2899 ¡36 9 12
−5794 35 96
– –
114 918 48 9 16
V-Clamp Im (pA) C (% D) a b
Expression phase
420 9 56 74 917
– –
– –
CNQX or DNQX present only during the trigger (induction) phase together with AMPA. CNQX present during the entire 120-min period.
30 min. Upon release of the current clamp, the Vms of PCs immediately returned to near control values (− 55 to − 65 mV), firing of action potentials slowed to a steady rhythmic pattern and membrane resistance increased slightly. This activity state was maintained for the remaining 90 min of the experimental protocol. Responses to microapplication of AMPA at this time point were identical to those obtained before electrical depolarization. Thus, sustained electrical depolarization of PCs for a period equivalent to, and exceeding, that produced by AMPA did not produce any enduring changes in PC electrophysiology.
4. Discussion Original studies by Garthwaite and Garthwaite (1990, 1991a) demonstrated excitotoxic DCD of hippocampal pyramidal neurons and cerebellar PCs exposed to AMPA for 30 min and subsequently observed 90 min after AMPA removal. Because the majority of PCs exhibit evidence of DCD, it is reasonable to assume that the majority of PCs exposed to AMPA and studied for electrophysiological changes in analogous experiments would also express DCD. Exposure of cerebellar slices to AMPA for 30 min elicited a characteristic profile of electrophysiologic activity during exposure to this agonist and for the subsequent 90-min period following AMPA removal. During exposure, AMPA produced a depolarization (inward current) with a corresponding decrease in Rin that diminished during the latter minutes of AMPA exposure. After AMPA removal, a sustained membrane depolarization (and underlying persistent inward current), an abundance of apparent miniature excitatory events, and loss of electro- and chemoresponsiveness to externally supplied stimuli occurred 60 – 75 min into this expression period. These changes potentially signify a severely compromised state in which PC viability was apparently low. The initial and secondary depolarizing events with accompanying conductance changes corre-
spond temporally to the induction and development of DCD in PCs and thus may represent an electrophysiologic profile of AMPA-induced trigger and expression of DCD. This profile is unlikely due to artifact arising from the whole-cell recording configuration and disruption of the PC intracellular milieu, because control recordings for an equivalent duration in the absence of AMPA showed no change in basal electrophysiology or chemoresponsiveness of PCs. Furthermore, because the membrane resistance remained relatively unchanged throughout the concomitant depolarization and reduced membrane responsiveness of the expression period, a general breakdown in membrane integrity probably was not occurring. A sudden decrease in membrane resistance would be indicative of a loss in membrane integrity typically observed in cell lysis. Potentially the marked suppression of ligand- and voltagegated currents reflects a protective mechanism for conservation of energy from which the cell may recover over time. This seems unlikely though, in that AMPA produces DCD after 90 min in the majority (60–75%) of PCs, and this proportion does not decrease with increased expression times extending to 180 min (Garthwaite and Garthwaite, 1991a). The immediate depolarization and inward current with accompanying changes in membrane resistance and conductance declined to about 25% by 15 min into the AMPA exposure. This most likely reflects strong AMPA receptor desensitization well known to occur with certain subtypes of AMPA receptors with even very transient exposures of agonist (Zorumski and Thio, 1992). This is significant because desensitization directly limits influx of the cations Ca2 + and Na + and secondarily limits their influx through voltage-gated channels that are no longer activated. After the removal of AMPA, a sustained depolarization occurred throughout the expression phase. Although the depolarization could be secondary to a persistent change in ionic gradients, we favor the notion that the depolarization is dependent on changes in membrane properties. For example, the membrane resistance decreased coin-
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
cident with depolarization indicating increased current flux. Moreover, a stimulus that could have promoted ion redistribution such as manually setting the membrane potential to an equivalent depolarization as seen during AMPA superfusion, failed to induce a secondary sustained depolarization when the clamp was released. These findings coupled with other studies in which CNQX blocked AMPA-induced DCD when given only during the expression phase (a treatment not likely to affect ion redistribution) (Garthwaite and Garthwaite, 1991b; Strahlendorf et al., 1996) would indicate that the sustained depolarization is secondary to a change in membrane properties. A unique characteristic of processes leading to DCD is an apparent paradoxical sensitivity to AMPA receptor antagonists. DCD is not altered by antagonists present with AMPA during exposure but antagonists block DCD development if present after AMPA removal, i.e. during the expression period (Garthwaite and Garthwaite, 1991b). The presence of CNQX during the entire protocol, a procedure that has also previously been shown to block AMPA-induced DCD of PCs (Garthwaite and Garthwaite, 1991b), prevented the development of characteristic electrophysiological events during the expression phase. Specifically, PCs failed to exhibit the sustained depolarization and maintained electro- and chemoresponsiveness 60 – 75 min into the expression phase. However, the presence of AMPA antagonists only during the trigger phase did not prevent the delayed electrophysiologic events and loss of cellular responsiveness, although the magnitude of the AMPA-induced depolarization during the trigger phase was reduced by approximately half. Moreover, manually clamping the PC membrane to potentials similar to that obtained with AMPA and for equivalent times revealed no sustained change in PC physiology for the duration of the experimental protocol. This pattern of electrophysiologic transitions supports the notion that the trigger of DCD is not a process directly linked to ionotropic events or depolarization, such as is associated with NMDA-receptor mediated excitotoxicity (Choi et al., 1988; Lysko et al., 1989; Michaels and Rothman, 1990). The secondary depolarization may be characteristically more ionotropic because CNQX present at this time markedly reduced it. Thus, presumed non-ionotropic events initiated during AMPA exposure apparently trigger partial ionotropic processes during the expression period that culminate in development of DCD. These findings suggest a correlative relationship between delayed electrophysiological changes and morphological events of DCD, and may represent an electrophysiological signature of DCD. Identities of factors responsible for secondary electrophysiologic and toxic effects are unknown. In all likelihood, factors that are both intrinsic and extrinsic to PCs are involved because in culture systems which lack
161
the relatively intact neuropil of the slice preparation, AMPA and kainate elicit only edema and necrosis of PCs (Brorson et al., 1994). The characteristic appearance of miniature excitatory potentials during the expression period that preceded the loss of neuronal viability underscores the potential importance of an intact neuropil to elicit DCD. Electrophysiologic responses of cultured hippocampal pyramidal cells to prolonged exposure to glutamate have been studied and differ from AMPA-induced effects on PCs in a number of ways (Sombati et al., 1991). Following 10-min application of glutamate, pyramidal neurons exhibited a persistent depolarization lasting up to 4 h with an increase in membrane conductance, but retained electro- and chemoresponsiveness, appeared morphologically normal and excluded vital dyes. The extended depolarization was both NMDA-receptor and Ca2 + -dependent and could not be mimicked by electrical depolarization of the neuron. Because the neuron remained viable during the entire recording period, it was concluded that the glutamate-elicited extended (4 h) depolarization probably served as an induction or trigger of more delayed events, occurring after the recording period, ultimately leading to cell death. In acutely isolated hippocampal CA1 neurons, NMDA also produced a continuously increasing postexposure inward cationic current (Na + and Ca2 + influx) (Chen et al., 1997). In contrast, PCs repolarize during AMPA exposure, undergo a second spontaneous depolarization after AMPA removal, lose electro- and chemoresponsiveness, undergo cytosolic condensation (Garthwaite and Garthwaite, 1991a) and calcium overload from external sources is apparently minimally involved (Garthwaite and Garthwaite, 1991b; Strahlendorf et al., 1998). Thus, the time frame and physiology of AMPA-elicited DCD in PCs differs markedly from NMDA receptor-mediated events in hippocampal neurons which also can express DCD in response to AMPA-receptor activation. It is unknown whether an electrophysiologic profile more similar to that in PCs would be elicited in pyramidal neurons in response to AMPA rather than glutamate. It is evident that excitotoxic electrophysiologic changes are neuron and EAA receptor specific and potentially offer insight into differences in mechanisms of cell death. However, it is difficult to unequivocally assess whether electrophysiologic alterations represent the cause of neuron death or rather are the effects of other factors causing the demise of the neuron. In a separate set of experiments, we have studied AMPA-elicited changes in PC intracellular Ca2 + during the entire DCD protocol (Strahlendorf et al., 1998). Exposure to AMPA elicited two peak increases in intracellular Ca2 + at times exactly corresponding to the initial and secondary depolarizations observed in the current experiments. Moreover, the presence of CNQX
162
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
or the absence of extracellular Ca2 + during the 30-min AMPA exposure did not change the Ca2 + transient pattern, but if these interventions were maintained for the entire protocol or only during the 90-min expression period, the secondary Ca2 + transient was blocked. These same conditions are ineffective and effective, respectively, in blocking DCD. Collectively, the data from these experiments extend and support previous studies (Garthwaite et al., 1986; Garthwaite and Garthwaite, 1991a) that DCD is largely independent of massive Ca2 + influx and potentially involves small fluxes or release of local intracellular Ca2 + stores. These small Ca2 + transients potentially initiate an active cascade of metabolic reactions that lead to DCD as opposed to passive physical damage (swelling and lysis) commonly seen with cation influx-dependent excitotoxicity (Choi et al., 1988). In this regard, DCD closely resembles apoptosis. In in vivo preparations, non-NMDA receptor-mediated toxicity has been reported to induce an apoptotic-like toxicity that morphologically resembles DCD (Filipkowski et al., 1994; Pollard et al., 1994; Portera-Cailliau et al., 1997). We have recently obtained preliminary findings that DNA damage occurs with DCD and DCD can be attenuated by inhibitors of protein synthesis, further suggesting that DCD enlists active rather than passive mechanisms. In summary, transient exposure to low levels of AMPA known to produce delayed dark cell degeneration of cerebellar Purkinje cells in the slice preparation, produced a characteristic electrophysiological response pattern that rendered the cell incapable of generating spontaneous or evoked electrical and chemical activity. This may have relevance to mechanisms underlying slow neurodegenerative processes.
Acknowledgements This work was supported by grants from Whitehall Foundation, National Ataxia Foundation and Texas Advanced Research Program.
References Audinat, E., Knopfel, T., Gahwiler, B.H., 1990. Responses to excitatory amino acids of Purkinje cells and neurones of the deep nuclei in cerebellar slice cultures. J. Physiol. 430, 297–313. Beal, M.F., 1992. Mechanisms of excitotoxicity in neurologic diseases. FASEB J. 6, 3338 – 3344. Brorson, J.R., Manzolillo, P.A., Miller, R.J., 1994. Ca2 + entry via AMPA/KA receptors and excitotoxicity in cultured cerebellar Purkinje cells. J. Neurosci. 14, 187–197. Chen, Q.X., Perkins, K.L., Choi, D.W., Wong, R.K.S., 1997. Secondary activation of a cation conductance is responsible for NDMA toxicity in acutely isolated hippocampal neurons. J. Neurosci. 17, 4032 – 4036.
Choi, D.W., Koh, J., Peters, S., 1988. Pharmacology of glutamate neurotoxicity in cortical cell culture: attenuation by NMDA antagonists. J. Neurosci. 8, 185 – 196. Filipkowski, R.K., Hetman, M., Kaminska, B., Kaczmarek, L., 1994. DNA fragmentation in rat brain after intraperitoneal administration of kainate. Neuroreport 5, 1538 – 1540. Garthwaite, J., Garthwaite, G., 1990. Excitatory amino acids and neuronal plasticity. In: Ben-Ari, Y. (Ed.), Mechanisms of Excitatory Amino Acid Neurotoxicity in Rat Brain Slices. Plenum Press, New York, pp. 505 – 518. Garthwaite, G., Garthwaite, J., 1991. AMPA neurotoxicity in rat cerebellar and hippocampal slices: histological evidence for three mechanisms. Eur. J. Neurosci. 3, 715 – 728. Garthwaite, G., Garthwaite, J., 1991. Mechanisms of AMPA neurotoxicity in rat brain slices. Eur. J. Neurosci. 3, 729 – 736. Garthwaite, J., Garthwaite, G., Hajos, F., 1986. Amino acid neurotoxicity: relationship to neuronal depolarization in rat cerebellar slices. Neuroscience 18, 449 – 460. Krupa, M., Crepel, F., 1990. Transient sensitivity of rat cerebellar Purkinje cells to N-methyl-D-asparate during development. A voltage clamp study in in vitro slices. Eur. J. Neurosci. 2, 312–316. Llano, I., Marty, A., Johnson, J.W., Ascher, P., Gahwiler, B.H., 1988. Patch-clamp recording of amino acid-activated responses in ‘organotypic’ slice cultures. Proc. Natl. Acad. Sci. USA 85, 3221– 3225. Llano, I., Marty, A., Armstrong, C.M., Konnerth, A., 1991. Synapticand agonist-induced excitatory currents of Purkinje cells in rat cerebellar slices. J. Physiol. 434, 183 – 213. Lysko, P.G., Cox, J.A., Vigano, M.A., Henneberry, R.C., 1989. Excitatory amino acid neurotoxicity at the N-methyl-D-asparate receptor in cultured neurons: pharmacological characterization. Brain Res. 499, 258 – 266. Meldrum, B., Garthwaite, J., 1990. Excitatory amino acid neurotoxicity and neurodegenerative disease. Trends Pharmacol. Sci. 11, 379 – 387. Michaels, R.L., Rothman, S.M., 1990. Glutamate neurotoxicity in vitro: antagonist pharmacology and intracellular calcium concentrations. J. Neurosci. 10, 283 – 292. Nellgard, B., Wieloch, T.J., 1992. Postischemic blockade of AMPA but not NMDA receptors mitigates neuronal damage in the rat brain following transient severe cerebral ischemia. Cereb. Blood Flow Metab. 12, 2 – 11. Pollard, H., Charriaut-Marlangue, C., Cantagrel, S., Represa, A., Robain, O., Moreau, J., Ben-Ari, Y., 1994. Kainate-induced apoptotic cell death in hippocampal neurons. Neuroscience 63, 7 – 18. Portera-Cailliau, C., Price, D.L., Martin, L.J., 1997. Non-NMDA and NMDA receptor-mediated excitotoxic neuronal deaths in adult brain are morphologically distinct: further evidence for an apoptosis – necrosis continuum. J. Comp. Neurol. 378, 88 – 104. Rosenmund, C., Legendre, P., Westbrook, G.L., 1992. Expression of NMDA channels on cerebellar Purkinje cells acutely dissociated from newborn rats. J. Neurophysiol. 68, 1901 – 1905. Sombati, S., Coulter, D.A., DeLorenzo, R.J., 1991. Neurotoxic activation of glutamate receptors induces an extended neuronal depolarization in cultured hippocampal neurons. Brain Res. 566, 316–319. Strahlendorf, J.C., Acosta, S., Strahlendorf, H.K., 1996. Diazoxide and cyclothiazide convert AMPA-induced dark cell degeneration of Purkinje cells to edematous damage in the cerebellar slice. Brain Res. 729, 197 – 204. Strahlendorf, J.C., Brandon, T., Miles, R., Strahlendorf, H.K., 1998. AMPA receptor-mediated alterations of intracellular calcium homeostasis in rat cerebellar Purkinje cells in vitro: Correlates to dark cell degeneration. Neurochem. Res. 23, 1355 – 1362. Zorumski, C.F., Thio, L.L., 1992. Properties of vertebrate glutamate receptors: calcium mobilization and desensitization. Prog. Neurobiol. 39, 295 – 336.
Enduring changes in Purkinje cell electrophysiology following transient exposure to AMPA: correlates to dark cell degeneration Jean C. Strahlendorf a,*, Howard K. Strahlendorf b b
a Texas Tech Uni6ersity, Health Science Center, Department of Physiology, 3601 4th St., Lubbock, TX 79430, USA Texas Tech Uni6ersity, Health Science Center, Department of Pharmacology, 3601 4th St., Lubbock, TX 79430, USA
Received 3 August 1998; accepted 11 December 1998
Abstract Purkinje cells (PCs) are selectively vulnerable to a-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA)-mediated delayed toxicity that is manifested as dark cell degeneration (DCD) rather than necrosis. The purpose of the present study was to utilize electrophysiologic changes induced by AMPA to gain mechanistic insights into its cytotoxic actions. The whole-cell configuration of the patch clamp technique was used to record spontaneous electrical activity and ionic currents of Purkinje neurons from cerebellar slices using an experimental paradigm known to produce DCD in response to AMPA. Initial electrophysiologic responses to AMPA consisted of a large transient depolarization and inward current that declined by 75% 20 min into the 30-min exposure to 30 mM AMPA. Cellular responses temporarily continued towards basal levels following removal of AMPA. A sustained membrane depolarization (and underlying persistent inward current), an abundance of apparent excitatory synaptic events, and loss of electro- and chemoresponsiveness were observed 60 – 75 min into the expression phase (following AMPA removal). These events correspond temporally to the development of DCD in Purkinje cells and may represent an electrophysiological signature of AMPA receptor-mediated delayed neurotoxic events. Antagonists of the AMPA receptor present concomitantly with AMPA are known not to affect DCD and failed to alter the electrophysiologic changes. The secondary depolarization and loss of electroresponsiveness were prevented by antagonists present after removal of AMPA, at a time when DCD also is prevented. Electrical clamping of the PC membrane to equivalent depolarized membrane potentials (Vms) obtained with AMPA failed to elicit any long lasting alterations in PC physiology. Collectively, morphological and electrophysiological data indicate that induction of DCD is not strongly dependent on ionotropic mechanisms elicited by AMPA receptors, but that expression of DCD does possess an ionotropic element. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: AMPA; Glutamate; Excitatory amino acids; Purkinje cells; Neurotoxicity; Electrophysiology
1. Introduction Although glutamate (glu) is benign as an excitatory transmitter in normal concentrations, abnormal glutamatergic tone may play a role in both acute neurotoxicity and chronic neurodegenerative processes. A sudden massive glu release has been linked intimately to acute neurotoxicity associated with cerebral ischemia, * Corresponding author. Tel.: +1-806-743-2554; fax: + 1-806-7431512.; e-mail: [email protected].
seizures, hypoglycemia, and trauma (Meldrum and Garthwaite, 1990). In contrast, mechanisms of chronic neurodegenerative diseases including hereditary ataxias are less understood but may be related to slow excitotoxic processes initiated inter alia by an impairment of energy metabolism, excitatory amino acid (EAA) receptor abnormalities, and slow or periodic excessive glutamatergic tone (Beal, 1992). It is now clear that excitotoxicity is multidimensional involving multiple glu receptor subtypes and mechanisms that produce several neurotoxic profiles. A wide-
0168-0102/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 8 - 0 1 0 2 ( 9 8 ) 0 0 1 2 6 - 6
156
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
spread salient role for AMPA receptors in delayed or latent excitotoxicity that equals or exceeds the role of NMDA receptors has emerged recently. Indeed, AMPA receptor-induced degeneration most often elicits a delayed toxicity (Garthwaite and Garthwaite, 1990; Nellgard and Wieloch, 1992), and unlike the widely studied mechanisms underlying acute edematous necrosis (EN) often mediated by abusive NMDA receptor activation, cellular processes associated with delayed degeneration remain relatively obscure. Minimally, delayed neuronal death is a multistage process in which transient exposure to high concentrations of EAAs serves as an initiation, induction or trigger event that promotes maturation of secondary cytotoxic cascades (Meldrum and Garthwaite, 1990). Thus, in delayed neurotoxicity, a striking temporal decoupling of toxic exposure from death is evident. Purkinje cells (PCs) express a paucity of functional NMDA receptors but a preponderance of AMPA receptors (Llano et al., 1988, 1991; Audinat et al., 1990; Krupa and Crepel, 1990; Rosenmund et al., 1992) and they are selectively vulnerable to delayed AMPA receptor-elicited toxicity (Garthwaite and Garthwaite, 1990). Therefore, the PC represents a unique neuron to study the cellular processes underlying AMPA toxicity in relative isolation of NMDA-mediated events. Delayed AMPA-induced toxicity has been demonstrated in cerebellar and hippocampal slices (Garthwaite and Garthwaite, 1991a) where AMPA was a potent neurotoxin capable of eliciting dark cell neurodegeneration (DCD) of PCs and pyramidal neurons hours after a brief exposure. AMPA-induced delayed DCD of PCs was found to be independent of both extracellular Ca2 + and AMPA antagonists, if present concomitantly with AMPA exposure; however antagonists did prevent AMPA-induced DCD when introduced after AMPA was removed (Garthwaite and Garthwaite, 1991b). This implies that an enduring AMPA receptor activation in PCs may be responsible for DCD. Electrophysiologic recordings of a population response from cerebellar cortical neurons indicated depolarization during the expression phase following AMPA removal (Garthwaite et al., 1986). Collectively, these findings indicate ionotropism is of little importance to the induction of AMPA-induced DCD in PCs, but may play a role in its expression. Although the studies of Garthwaite and colleagues (Garthwaite et al., 1986; Garthwaite and Garthwaite, 1990; Meldrum and Garthwaite, 1990) provided seminal data regarding AMPA-induced degeneration of PCs, these studies have not addressed the neurophysiology associated with the AMPA-selective delayed, degenerative process at the single PC level, the major target neuron. Changes in electrical properties of a neuron can provide a sensitive, immediate, and continuous index of the neuron’s physiology that can reflect
processes underlying cell death. Accordingly, the purpose of the present study was to gain insight into potential mechanisms underlying DCD by studying the electrophysiology of PCs subjected to AMPA receptormediated DCD.
2. Materials and methods Postnatal 8- to 12-day-old rats (Sprague–Dawley strain, Sasco Labs) were used in these studies. Sagittal slices (400 mm thick) of the vermal region of the cerebellum were obtained with a Vibroslice (Campden Instruments). Slices were allowed to recover for 1 h at room temperature submerged in artificial cerebrospinal fluid (ACSF) containing the following (in mM): NaCl (120), KCl (2), KH2PO4 (1.2), MgSO4 (1.2), CaCl2 (2), NaHCO3 (26), and glucose (11). For electrophysiologic recordings, slices were placed in an interface recording chamber, semi-submerged in ACSF flowing under and around them (1–3 ml/min) and blanketed with humidified 95:5 (%) O2/CO2. Slices were subjected to an experimental protocol that consisted of a 30-min treatment with 30 mM AMPA (termed trigger or induction phase), followed by a 90-min period in AMPA-free ACSF (expression phase). Control slices were exposed to the same protocol but in the absence of AMPA for the first 30 min. This experimental paradigm has reliably produced DCD in the majority of PCs (Garthwaite and Garthwaite, 1991a,b; Strahlendorf et al. 1996). Antagonists were superfused 10 min prior to the application of AMPA and given for the entire protocol or only during the trigger (induction) phase. The whole-cell configuration of the patch clamp technique was used to record spontaneous electrical activity under current-clamp and ionic currents under voltageclamp. Glass tubing blanks were pulled to produce electrodes with a tip resistance of 3–8 MV. The electrode solution contained (in mM): 140 K + gluconate, 2.5 MgCl2, 10 HEPES, 0.6 EGTA, 0.06 CaCl2, 4 NaATP, and 0.4 Na-GTP, pH 7.2. On occasion, Lucifer Yellow (1 mg/ml) was included in the solution to label the recorded cell for identification after the experiment. Electrical signals were led into a Dagan 3900 amplifier and displayed on a storage oscilloscope. Signals were permanently recorded on videotape using an analog to digital converter (PCM-2 A/D, Medical Systems) connected to a video recorder (Panasonic). An IBM-compatible AT computer equipped with an Axon Instruments TL-1 interface and pClamp software (Version 5.5) was used for data acquisition and analysis. Electrical recordings from PCs were made at 33°C. Only healthy, viable PCs were used to assess AMPA-induced toxicity. Viability was assessed using the following criteria: (1) a stable resting potential of at least − 55 mV for greater than 10 min, (2) the ability to fire
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
157
Fig. 1. Whole-cell current clamp recording of an individual Purkinje cell (PC) in the cerebellar slice preparation in response to AMPA (30 mM) superfused for 30 min. (A) PC membrane response during AMPA exposure (beginning at the first downward arrowhead and ending at the second downward arrowhead). AMPA elicited a depolarization with continuous Na + and Ca2 + spikes (denoted by an unfilled circle) until the membrane depolarized to a potential of approximately − 23 mV. This was accompanied by a decrease in input resistance. A sustained depolarization (indicated by an asterisk) peaked near − 18 mV. During the latter portion of the AMPA exposure period, the membrane potential remained depolarized at approximately − 30 mV and Rin began to return towards control levels. (B) PC membrane response immediately after AMPA exposure (0 – 30 min into expression phase). The PC maintained a sustained depolarization that ranged between − 25 and − 40 mV with a decrease in resistance observed during the maximal depolarization. (C) PC membrane response 30 – 60 min into the expression phase. The membrane had returned towards basal potential and Rin before initiating an increase in electrical activity. The PC membrane continued to exhibit an increase in spontaneous excitatory events with an accompanying depolarization towards − 30 mV (denoted by a crossed diamond) with a variable change in Rin. (D) PC response 60–90 min into the expression phase.
spontaneous or current-driven action potentials, (3) a reasonably high membrane resistance (\ 200 MV), and (4) the presence of hyperpolarization-activated inward current (Ih), since this current is present in all healthy PCs. Pipette capacitance was neutralized to a minimum before rupture; after rupture, the height of the capacitive transient in response to a 10-mV test pulse was used to calculate and compensate (\75%) the series resistance. Changes in magnitudes of membrane potentials (Vms), membrane currents, slope resistances, slope conductances, and electro- and chemoresponsiveness were assessed after AMPA applications. Slope resistances and conductances were calculated from I/V or V/I curves generated by 400 ms currents (10 pA incrementally) or voltage pulses (10 mV incrementally), respectively. Electroresponsiveness was evaluated at regular intervals by the capability of PCs to generate voltagedependent active membrane responses from − 60 mV. Specifically, the ability of PCs to fire action potentials or generate inward current in response to depolarizing current or voltage steps, respectively, and the capacity of PCs to elicit Ih in response to hyperpolarizing pulses was analyzed. Maintenance of chemoresponsiveness was measured by the magnitude of membrane responses (mV or pA) elicited by local AMPA applications of 10-ml microdrops, 75 mM. Loss of functional viability was defined as the inability to fire action potentials upon depolarization and loss of ligand- and/ or voltage-gated currents. Experiments in which recordings of Vm and input resistance (Rin) were unstable, series resistance could not be adequately compensated, or a sudden loss in Vm, Rin, and electro- and chemore-
sponsiveness occurred (most likely due to rupture of the plasma membrane), were not included in the data analysis. All electrophysiological parameters were normalized to pre-AMPA conditions, and post-AMPA changes expressed as percent differences. The data were expressed as the mean percentage9 S.E.M.
3. Results Control studies without AMPA superfusion (n=3) were done to establish the long-term stability of the recording procedure. These experiments revealed excellent preservation of PC electro- and chemoresponsiveness for more than 3 h. Specifically, PCs maintained membrane potentials close to − 60 mV without loss of monitored Na + - and Ca2 + -dependent action potentials and voltage-gated currents such as Ih. Additionally, PCs maintained chemical responsiveness to periodic local microapplications of AMPA, demonstrating reproducible depolarizations throughout a 3 h recording period. Loss of whole cell recordings marked by a sudden loss of Rin and a marked depolarizing shift in Vm to near 0 mV that failed to return towards control values occasionally occurred. Data from these PCs were excluded from further analysis. In current clamp mode, AMPA (30 mM, 30 min superfusion, n=13) elicited in PCs a characteristic pattern of membrane responses (Fig. 1). In all PCs during the first 5–10 min, AMPA elicited a depolarization with continuous composite Na + and Ca2 + action potentials until the membrane reached a potential that
158
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
averaged −25 ( 92 S.E.M.) mV (Fig. 1A). Subsequently, a transient 5 mV repolarization lasting approximately 15 s preceded a sustained depolarization that peaked near −10 (94) mV. The latter depolarization was accompanied by an average 31(9 6)% decrease in Rin. During the latter 10 min of the AMPA exposure period, membrane resistance returned to near control (pre-AMPA) levels, although Vm remained depolarized to an average plateau potential of − 40 (9 4) mV. During the AMPA expression period (time during which DCD is expressed), PCs remained depolarized to an average of −38 (9 5) mV, although Rin remained within 20% of the control levels (Fig. 1B). Apparent Ca2 + action potentials and compound miniature excitatory potentials (10 mV in amplitude and 15 ms in duration) appeared 40 – 50 min into the expression period (Fig. 1C). Typically, 60 – 75 min into the expression period, the Vm showed a continuously increasing sustained depolarization towards − 20 ( 9 6) mV with a maximal decrease in Rin (average 42911%). The depolarization prevailed until terminating the experiment at 90 min (Fig. 1D). During this period, manually clamping the Vm back to control values ( −55 to −60 mV) also revealed loss of spontaneous and depolarizationevoked Ca2 + - and Na + -dependent action potentials, and disappearance of voltage-gated responses (Ih) to hyperpolarizing current pulses. Prior to the 30-min AMPA exposure period, local microapplication of AMPA (75 mM, 10 ml) onto PCs produced an average 35 ( 97) mV depolarization with an average 13 (93)% decrease in Rin followed by a 6 ( 92) mV hyperpolarization and a slight increase in Rin (average 15 9 6%). Responsiveness of PCs to local microapplications of AMPA was markedly reduced by 60 – 75 min into the expression phase and at the end of the experimental protocol (i.e. at the end of the 90-min expression phase). Specifically, AMPA elicited only a 5 ( 9 2) mV depolarization with no apparent accompanying membrane resistance change or secondary hyperpolarization with the Vm manually clamped at control values. Voltage-clamp recordings revealed a characteristic profile of membrane currents in response to AMPA. During AMPA application, PCs exhibited a marked, sustained inward current that averaged 420 (956, n= 4) pA with an accompanying increase in membrane conductance averaging 74 (9 17)% of control that returned towards control values during the latter minutes of the exposure period (Fig. 2A). During the ensuing 60 min following AMPA exposure, membrane current and conductance migrated towards control levels and spontaneous, unclamped regenerative currents began to appear at times that varied between 10 and 30 min (Fig. 2B,C). A sustained inward current (average 114 918 pA) and increase in conductance (average 48916% of control) occurred approximately 60 min into the ex-
pression phase and temporally coincided with the sustained depolarization and decreased membrane resistance observed under current-clamp conditions (Fig. 2D). In the majority of PCs studied, a flurry of inward compound miniature excitatory currents preceded or occurred coincident with the onset of the sustained inward current (Fig. 2D). Loss of electroresponsiveness 60–75 min into the expression phase was also evident by failure to elicit in PCs Ih, ICa2 + or INa + following application of appropriate voltage steps from a Vm clamped at − 60 mV. To assess the contribution of ionotropic processes in mediating AMPA-induced electrophysiological events, the competitive antagonist 6-cyano-7-nitroquinoxaline2,3-dione (CNQX) (10 mM, n= 3) was superfused during the entire protocol (with AMPA and during the expression phase). This paradigm of CNQX application
Fig. 2. Whole-cell voltage clamp recording of an individual Purkinje cell (PC) in the cerebellar slice preparation in response to AMPA (30 mM) superfused for 30 min. (A) A 30-min AMPA exposure (beginning at the diamond symbol) elicited a double-peaked inward current with a transient flurry of unclamped regenerative currents (indicated by the arrowhead) and an increase in membrane conductance. During the latter minutes of AMPA exposure, a sustained inward current of approximately 150 pA was evident. (B) PC membrane current response immediately after AMPA exposure (0 – 30 min into expression phase). The PC maintained its pre-AMPA current values that ranged between 50 and 100 pA with spontaneous unclamped regenerative currents. Occasional inward plateau currents elicited bursts of regenerative currents that spontaneously subsided (denoted by an asterisk). (C) PC membrane response 30 – 60 min into the expression phase. The inward current continued at basal levels with occasional spontaneous regenerative currents and membrane conductance also had recovered to near control, pre-AMPA, values. Amplitudes of unclamped regenerative currents typically began to diminish 50 – 60 min into the expression period. (D) PC currents 60 – 90 min into the expression phase. The PC membrane continued to exhibit an increase in spontaneous excitatory regenerative currents (indicated by an asterisk) that culminated in a prolonged inward current shift of approximately 200 pA (indicated by a crossed diamond) and increased membrane conductance. This condition persisted until termination of the recording at greater than 90 min after AMPA exposure.
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
159
Fig. 3. Response of a PC to AMPA (30 mM) superfused for 30 min in the continuous presence of CNQX 10 mM using whole-cell current clamp recording. CNQX superfusion was started 10 min prior to application of AMPA. (A) In the presence of CNQX, the PC displayed a marked depolarization that peaked near −16 mV, with an associated decrease in Rin. However, unlike the response observed with AMPA only, Rin and Vm rapidly returned to near control values accompanied with continuous Na + and Ca2 + spikes within 5 min after the onset of AMPA application. Diamond denotes the beginning of the AMPA application and the bar represents the basal Vm. (B) Response of the same PC, 35–45 min into the AMPA expression phase (following AMPA removal, CNQX still present). The PC continued to spontaneously fire action potentials with a Vm slightly depolarized at approximately −53 mV and a modest increase in Rin (approximately 25% increase). (C) The response of the PC 60 – 75 min into the expression phase. As is evident, the characteristic secondary depolarization observed during the expression phase was absent when CNQX was present. (D) The response of the PC 90 min into the expression phase. Vm was approximately −60 mV with an approximate 25% increase in Rin. The PC continued to exhibit synaptic potentials and spontaneous action potentials.
previously was reported to effectively block AMPA-induced DCD (Garthwaite and Garthwaite, 1991a). In the presence of CNQX, AMPA still exhibited a marked depolarization of PCs that peaked at an average − 27 ( 9 5) mV compared to − 10 mV in the absence of CNQX (Fig. 3A). Although the associated decrease in Rin was equivalent initially to that seen in the absence of CNQX (31[ 98] vs 31[ 96]%, respectively), Rin rapidly returned to near pre-AMPA values 5 min into the AMPA application. In the presence of CNQX, the AMPA-induced plateau potential reached an average of − 51 (9 3) mV compared to −40 (9 4) mV in the absence of CNQX. In the presence of CNQX during the ensuing 60 min following the removal of AMPA, the membrane potential of PCs and accompanying input resistance remained at essentially control levels, at most evincing a 5-mV depolarization (Fig. 3B,C). The secondary depolarization observed during the expression phase was absent with CNQX present (Fig. 3D). Moreover, CNQX preserved the ability of PCs to continuously discharge spontaneous Ca2 + and Na + action potentials, generate voltage-dependent membrane responses (Ih) and display miniature excitatory potentials. Furthermore, CNQX maintained Rin at a higher level (420 MV AMPA + CNQX vs 310 MV AMPA alone) indicating that the membrane was not excessively leaky (Fig. 3B–D). To determine the contribution of ionotropic processes during the trigger phase, competitive AMPA receptor antagonists CNQX (10 mM, n =3, IC50 = 0.3 mM) or DNQX (30 and 60 mM, n = 5, IC50 =0.5 mM) were superfused with AMPA only during the 30-min
trigger phase, and then removed. Characteristic electrophysiologic events observed during the expression phase (see above) were minimally affected by the presence of either of the two antagonists at any of the given concentrations during the trigger phase (data not shown). This was not unexpected given the similarity in IC50s and previously reported morphological experiments that demonstrated conclusively that AMPA receptor antagonists present only during the trigger phase failed to block DCD (Garthwaite and Garthwaite, 1991a). Analogous to the expression phase in AMPA alone experiments, miniature excitatory potentials appeared within 15 min and increased in frequency during the following 50–60 min until regenerative spikes appeared, superimposed on a depolarizing Vm. A sustained depolarization to an average − 28 ( 9 9) mV (with an average 36 [9 12]% decrease in Rin) peaked 75 min into the expression phase. Moreover, voltage-gated currents (INa and Ih) were also blocked totally during the latter stages of the expression period. Thus, inclusion of an ionotropic EAA antagonist coincident with AMPA, but not during the expression phase, failed to prevent the characteristic delayed and sustained excitatory events during the expression period seen following application of only AMPA. A summary of electrophysiological changes observed in PCs exposed to AMPA alone and in the presence of antagonists is provided in Table 1. To ascertain whether only a 30-min sustained depolarization of the membrane alone could be sufficient to induce secondary electrophysiologic effects similar to AMPA exposure, PCs were manually clamped to Vms equivalent to those elicited by AMPA superfusion for
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
160
Table 1 Summary of changes in membrane properties of Purkinje neurons exposed to AMPA Trigger/induction phase
C-Clamp Vm (mV) Rin (% D)
AMPA
AMPA and ANTAG (trig)a
AMPA
AMPA and ANTAG (trig)a
AMPA and ANTAG (entire)b
−109 4 ¡31 96
−279 5 ¡3198
−209 6 ¡42 9 11
−2899 ¡36 9 12
−5794 35 96
– –
114 918 48 9 16
V-Clamp Im (pA) C (% D) a b
Expression phase
420 9 56 74 917
– –
– –
CNQX or DNQX present only during the trigger (induction) phase together with AMPA. CNQX present during the entire 120-min period.
30 min. Upon release of the current clamp, the Vms of PCs immediately returned to near control values (− 55 to − 65 mV), firing of action potentials slowed to a steady rhythmic pattern and membrane resistance increased slightly. This activity state was maintained for the remaining 90 min of the experimental protocol. Responses to microapplication of AMPA at this time point were identical to those obtained before electrical depolarization. Thus, sustained electrical depolarization of PCs for a period equivalent to, and exceeding, that produced by AMPA did not produce any enduring changes in PC electrophysiology.
4. Discussion Original studies by Garthwaite and Garthwaite (1990, 1991a) demonstrated excitotoxic DCD of hippocampal pyramidal neurons and cerebellar PCs exposed to AMPA for 30 min and subsequently observed 90 min after AMPA removal. Because the majority of PCs exhibit evidence of DCD, it is reasonable to assume that the majority of PCs exposed to AMPA and studied for electrophysiological changes in analogous experiments would also express DCD. Exposure of cerebellar slices to AMPA for 30 min elicited a characteristic profile of electrophysiologic activity during exposure to this agonist and for the subsequent 90-min period following AMPA removal. During exposure, AMPA produced a depolarization (inward current) with a corresponding decrease in Rin that diminished during the latter minutes of AMPA exposure. After AMPA removal, a sustained membrane depolarization (and underlying persistent inward current), an abundance of apparent miniature excitatory events, and loss of electro- and chemoresponsiveness to externally supplied stimuli occurred 60 – 75 min into this expression period. These changes potentially signify a severely compromised state in which PC viability was apparently low. The initial and secondary depolarizing events with accompanying conductance changes corre-
spond temporally to the induction and development of DCD in PCs and thus may represent an electrophysiologic profile of AMPA-induced trigger and expression of DCD. This profile is unlikely due to artifact arising from the whole-cell recording configuration and disruption of the PC intracellular milieu, because control recordings for an equivalent duration in the absence of AMPA showed no change in basal electrophysiology or chemoresponsiveness of PCs. Furthermore, because the membrane resistance remained relatively unchanged throughout the concomitant depolarization and reduced membrane responsiveness of the expression period, a general breakdown in membrane integrity probably was not occurring. A sudden decrease in membrane resistance would be indicative of a loss in membrane integrity typically observed in cell lysis. Potentially the marked suppression of ligand- and voltagegated currents reflects a protective mechanism for conservation of energy from which the cell may recover over time. This seems unlikely though, in that AMPA produces DCD after 90 min in the majority (60–75%) of PCs, and this proportion does not decrease with increased expression times extending to 180 min (Garthwaite and Garthwaite, 1991a). The immediate depolarization and inward current with accompanying changes in membrane resistance and conductance declined to about 25% by 15 min into the AMPA exposure. This most likely reflects strong AMPA receptor desensitization well known to occur with certain subtypes of AMPA receptors with even very transient exposures of agonist (Zorumski and Thio, 1992). This is significant because desensitization directly limits influx of the cations Ca2 + and Na + and secondarily limits their influx through voltage-gated channels that are no longer activated. After the removal of AMPA, a sustained depolarization occurred throughout the expression phase. Although the depolarization could be secondary to a persistent change in ionic gradients, we favor the notion that the depolarization is dependent on changes in membrane properties. For example, the membrane resistance decreased coin-
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
cident with depolarization indicating increased current flux. Moreover, a stimulus that could have promoted ion redistribution such as manually setting the membrane potential to an equivalent depolarization as seen during AMPA superfusion, failed to induce a secondary sustained depolarization when the clamp was released. These findings coupled with other studies in which CNQX blocked AMPA-induced DCD when given only during the expression phase (a treatment not likely to affect ion redistribution) (Garthwaite and Garthwaite, 1991b; Strahlendorf et al., 1996) would indicate that the sustained depolarization is secondary to a change in membrane properties. A unique characteristic of processes leading to DCD is an apparent paradoxical sensitivity to AMPA receptor antagonists. DCD is not altered by antagonists present with AMPA during exposure but antagonists block DCD development if present after AMPA removal, i.e. during the expression period (Garthwaite and Garthwaite, 1991b). The presence of CNQX during the entire protocol, a procedure that has also previously been shown to block AMPA-induced DCD of PCs (Garthwaite and Garthwaite, 1991b), prevented the development of characteristic electrophysiological events during the expression phase. Specifically, PCs failed to exhibit the sustained depolarization and maintained electro- and chemoresponsiveness 60 – 75 min into the expression phase. However, the presence of AMPA antagonists only during the trigger phase did not prevent the delayed electrophysiologic events and loss of cellular responsiveness, although the magnitude of the AMPA-induced depolarization during the trigger phase was reduced by approximately half. Moreover, manually clamping the PC membrane to potentials similar to that obtained with AMPA and for equivalent times revealed no sustained change in PC physiology for the duration of the experimental protocol. This pattern of electrophysiologic transitions supports the notion that the trigger of DCD is not a process directly linked to ionotropic events or depolarization, such as is associated with NMDA-receptor mediated excitotoxicity (Choi et al., 1988; Lysko et al., 1989; Michaels and Rothman, 1990). The secondary depolarization may be characteristically more ionotropic because CNQX present at this time markedly reduced it. Thus, presumed non-ionotropic events initiated during AMPA exposure apparently trigger partial ionotropic processes during the expression period that culminate in development of DCD. These findings suggest a correlative relationship between delayed electrophysiological changes and morphological events of DCD, and may represent an electrophysiological signature of DCD. Identities of factors responsible for secondary electrophysiologic and toxic effects are unknown. In all likelihood, factors that are both intrinsic and extrinsic to PCs are involved because in culture systems which lack
161
the relatively intact neuropil of the slice preparation, AMPA and kainate elicit only edema and necrosis of PCs (Brorson et al., 1994). The characteristic appearance of miniature excitatory potentials during the expression period that preceded the loss of neuronal viability underscores the potential importance of an intact neuropil to elicit DCD. Electrophysiologic responses of cultured hippocampal pyramidal cells to prolonged exposure to glutamate have been studied and differ from AMPA-induced effects on PCs in a number of ways (Sombati et al., 1991). Following 10-min application of glutamate, pyramidal neurons exhibited a persistent depolarization lasting up to 4 h with an increase in membrane conductance, but retained electro- and chemoresponsiveness, appeared morphologically normal and excluded vital dyes. The extended depolarization was both NMDA-receptor and Ca2 + -dependent and could not be mimicked by electrical depolarization of the neuron. Because the neuron remained viable during the entire recording period, it was concluded that the glutamate-elicited extended (4 h) depolarization probably served as an induction or trigger of more delayed events, occurring after the recording period, ultimately leading to cell death. In acutely isolated hippocampal CA1 neurons, NMDA also produced a continuously increasing postexposure inward cationic current (Na + and Ca2 + influx) (Chen et al., 1997). In contrast, PCs repolarize during AMPA exposure, undergo a second spontaneous depolarization after AMPA removal, lose electro- and chemoresponsiveness, undergo cytosolic condensation (Garthwaite and Garthwaite, 1991a) and calcium overload from external sources is apparently minimally involved (Garthwaite and Garthwaite, 1991b; Strahlendorf et al., 1998). Thus, the time frame and physiology of AMPA-elicited DCD in PCs differs markedly from NMDA receptor-mediated events in hippocampal neurons which also can express DCD in response to AMPA-receptor activation. It is unknown whether an electrophysiologic profile more similar to that in PCs would be elicited in pyramidal neurons in response to AMPA rather than glutamate. It is evident that excitotoxic electrophysiologic changes are neuron and EAA receptor specific and potentially offer insight into differences in mechanisms of cell death. However, it is difficult to unequivocally assess whether electrophysiologic alterations represent the cause of neuron death or rather are the effects of other factors causing the demise of the neuron. In a separate set of experiments, we have studied AMPA-elicited changes in PC intracellular Ca2 + during the entire DCD protocol (Strahlendorf et al., 1998). Exposure to AMPA elicited two peak increases in intracellular Ca2 + at times exactly corresponding to the initial and secondary depolarizations observed in the current experiments. Moreover, the presence of CNQX
162
J.C. Strahlendorf, H.K. Strahlendorf / Neuroscience Research 33 (1999) 155–162
or the absence of extracellular Ca2 + during the 30-min AMPA exposure did not change the Ca2 + transient pattern, but if these interventions were maintained for the entire protocol or only during the 90-min expression period, the secondary Ca2 + transient was blocked. These same conditions are ineffective and effective, respectively, in blocking DCD. Collectively, the data from these experiments extend and support previous studies (Garthwaite et al., 1986; Garthwaite and Garthwaite, 1991a) that DCD is largely independent of massive Ca2 + influx and potentially involves small fluxes or release of local intracellular Ca2 + stores. These small Ca2 + transients potentially initiate an active cascade of metabolic reactions that lead to DCD as opposed to passive physical damage (swelling and lysis) commonly seen with cation influx-dependent excitotoxicity (Choi et al., 1988). In this regard, DCD closely resembles apoptosis. In in vivo preparations, non-NMDA receptor-mediated toxicity has been reported to induce an apoptotic-like toxicity that morphologically resembles DCD (Filipkowski et al., 1994; Pollard et al., 1994; Portera-Cailliau et al., 1997). We have recently obtained preliminary findings that DNA damage occurs with DCD and DCD can be attenuated by inhibitors of protein synthesis, further suggesting that DCD enlists active rather than passive mechanisms. In summary, transient exposure to low levels of AMPA known to produce delayed dark cell degeneration of cerebellar Purkinje cells in the slice preparation, produced a characteristic electrophysiological response pattern that rendered the cell incapable of generating spontaneous or evoked electrical and chemical activity. This may have relevance to mechanisms underlying slow neurodegenerative processes.
Acknowledgements This work was supported by grants from Whitehall Foundation, National Ataxia Foundation and Texas Advanced Research Program.
References Audinat, E., Knopfel, T., Gahwiler, B.H., 1990. Responses to excitatory amino acids of Purkinje cells and neurones of the deep nuclei in cerebellar slice cultures. J. Physiol. 430, 297–313. Beal, M.F., 1992. Mechanisms of excitotoxicity in neurologic diseases. FASEB J. 6, 3338 – 3344. Brorson, J.R., Manzolillo, P.A., Miller, R.J., 1994. Ca2 + entry via AMPA/KA receptors and excitotoxicity in cultured cerebellar Purkinje cells. J. Neurosci. 14, 187–197. Chen, Q.X., Perkins, K.L., Choi, D.W., Wong, R.K.S., 1997. Secondary activation of a cation conductance is responsible for NDMA toxicity in acutely isolated hippocampal neurons. J. Neurosci. 17, 4032 – 4036.
Choi, D.W., Koh, J., Peters, S., 1988. Pharmacology of glutamate neurotoxicity in cortical cell culture: attenuation by NMDA antagonists. J. Neurosci. 8, 185 – 196. Filipkowski, R.K., Hetman, M., Kaminska, B., Kaczmarek, L., 1994. DNA fragmentation in rat brain after intraperitoneal administration of kainate. Neuroreport 5, 1538 – 1540. Garthwaite, J., Garthwaite, G., 1990. Excitatory amino acids and neuronal plasticity. In: Ben-Ari, Y. (Ed.), Mechanisms of Excitatory Amino Acid Neurotoxicity in Rat Brain Slices. Plenum Press, New York, pp. 505 – 518. Garthwaite, G., Garthwaite, J., 1991. AMPA neurotoxicity in rat cerebellar and hippocampal slices: histological evidence for three mechanisms. Eur. J. Neurosci. 3, 715 – 728. Garthwaite, G., Garthwaite, J., 1991. Mechanisms of AMPA neurotoxicity in rat brain slices. Eur. J. Neurosci. 3, 729 – 736. Garthwaite, J., Garthwaite, G., Hajos, F., 1986. Amino acid neurotoxicity: relationship to neuronal depolarization in rat cerebellar slices. Neuroscience 18, 449 – 460. Krupa, M., Crepel, F., 1990. Transient sensitivity of rat cerebellar Purkinje cells to N-methyl-D-asparate during development. A voltage clamp study in in vitro slices. Eur. J. Neurosci. 2, 312–316. Llano, I., Marty, A., Johnson, J.W., Ascher, P., Gahwiler, B.H., 1988. Patch-clamp recording of amino acid-activated responses in ‘organotypic’ slice cultures. Proc. Natl. Acad. Sci. USA 85, 3221– 3225. Llano, I., Marty, A., Armstrong, C.M., Konnerth, A., 1991. Synapticand agonist-induced excitatory currents of Purkinje cells in rat cerebellar slices. J. Physiol. 434, 183 – 213. Lysko, P.G., Cox, J.A., Vigano, M.A., Henneberry, R.C., 1989. Excitatory amino acid neurotoxicity at the N-methyl-D-asparate receptor in cultured neurons: pharmacological characterization. Brain Res. 499, 258 – 266. Meldrum, B., Garthwaite, J., 1990. Excitatory amino acid neurotoxicity and neurodegenerative disease. Trends Pharmacol. Sci. 11, 379 – 387. Michaels, R.L., Rothman, S.M., 1990. Glutamate neurotoxicity in vitro: antagonist pharmacology and intracellular calcium concentrations. J. Neurosci. 10, 283 – 292. Nellgard, B., Wieloch, T.J., 1992. Postischemic blockade of AMPA but not NMDA receptors mitigates neuronal damage in the rat brain following transient severe cerebral ischemia. Cereb. Blood Flow Metab. 12, 2 – 11. Pollard, H., Charriaut-Marlangue, C., Cantagrel, S., Represa, A., Robain, O., Moreau, J., Ben-Ari, Y., 1994. Kainate-induced apoptotic cell death in hippocampal neurons. Neuroscience 63, 7 – 18. Portera-Cailliau, C., Price, D.L., Martin, L.J., 1997. Non-NMDA and NMDA receptor-mediated excitotoxic neuronal deaths in adult brain are morphologically distinct: further evidence for an apoptosis – necrosis continuum. J. Comp. Neurol. 378, 88 – 104. Rosenmund, C., Legendre, P., Westbrook, G.L., 1992. Expression of NMDA channels on cerebellar Purkinje cells acutely dissociated from newborn rats. J. Neurophysiol. 68, 1901 – 1905. Sombati, S., Coulter, D.A., DeLorenzo, R.J., 1991. Neurotoxic activation of glutamate receptors induces an extended neuronal depolarization in cultured hippocampal neurons. Brain Res. 566, 316–319. Strahlendorf, J.C., Acosta, S., Strahlendorf, H.K., 1996. Diazoxide and cyclothiazide convert AMPA-induced dark cell degeneration of Purkinje cells to edematous damage in the cerebellar slice. Brain Res. 729, 197 – 204. Strahlendorf, J.C., Brandon, T., Miles, R., Strahlendorf, H.K., 1998. AMPA receptor-mediated alterations of intracellular calcium homeostasis in rat cerebellar Purkinje cells in vitro: Correlates to dark cell degeneration. Neurochem. Res. 23, 1355 – 1362. Zorumski, C.F., Thio, L.L., 1992. Properties of vertebrate glutamate receptors: calcium mobilization and desensitization. Prog. Neurobiol. 39, 295 – 336.