JOURNALOF NEUROPHYSIOLOGY Vol. 65. No. 4, .4pril 199 1. Yrinfd
in
I’..S..d.
Epileptifom Activity in Microcultures Containing One Excitatory Hippocampal Neuron MICHAEL
M. SEGAL
Department of Neurology, Massachusetts General Hospital Harvard Medical School, Boston, Massachusetts 02115 SUMMARY
AND
CONCLUSIONS
1. Paroxysmal depolarizing shifts (PDSs) occur during interictal epileptiform activity. Sustained depolarizations are characteristic of ictal activity, and events resembling PDSs also occur during the sustained depolarizations. To study these elements of epileptiform activity in a simpler context, I used the in vitro chronic-excitatory-block model of epilepsy of Furshpan and Potter and the microculture technique of Segal and Furshpan. 2. Intracellular recordings were made from 93 single-neuron microcultures. Forty of these solitary neurons were excitatory; their action potentials were replaced by PDS-like events or sustained depolarizations as kynurenate was removed from the perfusion solution. PDS-like events were similar to PDSs in intact cortex, mass cultures, and microcultures with more than one neuron. Small voltage fluctuations were also seen in solitary excitatory neurons in the absence of recorded action potentials. Sustained depolarizations developed in 5 of the 40 excitatory neurons. Forty-eight of the 93 solitary neurons were inhibitory, with bicuculline-sensitive hyperpolarizations after the action potential (ascribable to y-aminobutyric acid-A autapses). None of the solitary inhibitory neurons displayed sustained depolarizations. Five of the 93 neurons were insensitive to both kynurenate and bicuculline and were not placed in either the excitatory or the inhibitory category. 3. Both N-methyl-D-aspartate (NMDA) and non-NMDA glutamate receptors contributed to the PDS-like events and sustained depolarizations. Only a non-NMDA glutamate receptor component was evident for the small voltage fluctuations. 4. Intracellular recordings were also made from two-neuron microcultures, each containing one excitatory neuron and one inhibitory neuron. Sustained depolarizations developed in five microcultures, in each case only in the excitatory neuron. INTRODUCTION
and Department
of Neurobiology,
san 1964a,b; Rutecki et al. 1985). In contrast, sustained depolarizations are events that last seconds to minutes, and are associated with ictal activity (Hablitz 1987; Haglund and Schwartzkroin 1984; Kandel and Spencer 196 1; Matsumoto and Ajmone Marsan 1964b; Somjen et al. 1985). To study these elements of epileptiform activity, a potentially useful approach is to reduce the complexity of the system and examine the PDSs and sustained depolarizations in the simplest neuronal circuits in which they occur. A model of epilepsy has been described using mass cultures containing hundreds or thousands of neurons (Furshpan and Potter 1989). The cells are dissociated from the hippocampal formations of neonatal rats and are grown chronically in the presence of agents that suppress synaptic activity [elevated Mg2+ and kynurenate, a blocker of both N-methyl-D-aspartate (NMDA) and non-NMDA type glutamate receptors]. After several weeks, intense epileptiform activity can be recorded arising spontaneously when the blocking agents are removed. This activity displays many features of seizures experimentally induced in intact cortex: large PDS-like events; sudden increases in their frequency, often accompanied by sustained depolarizations; activation of inhibitory as well as excitatory mechanisms (Kandel and Spencer 196 1; Rutecki et al. 1985); “postictal” hyperpolarizations when the episodes are terminated (Kandel and Spencer 196 1); and extensive neuronal death when episodes are not terminated (Siesjo and Wieloch 1986). Microcultures containing one or several neurons have been used as a simple system to study the synaptic functions of sympathetic neurons (Furshpan et al. 1976, 1986). A microculture version of the mass culture epilepsy model has been described recently (Segal and Furshpan 1990). In this model, microcultures containing several neurons were grown chronically in the presence of the synaptic blockers kynurenate and elevated Mg2+. In microcultures containing 2- 18 neurons, on removal of the blockers, transitions to seizure-like activity and PDS-like events were observed that were similar to the epileptiform activity observed in mass cultures. I report here that both PDS-like events and sustained depolarizations can occur in microcultures that contain a single excitatory neuron. Some of these findings have been reported in preliminary form (Segal 1990).
Intracellular recordings from cortical neurons in intactanimal models of epilepsy have demonstrated that paroxysmal depolarizing shifts (PDSs) are characteristic of interictal epileptiform activity, and events resembling PDSs are also seen as part of the ictal activity. In vivo, PDSs are large, complex depolarizing events in which the cell rapidly depolarizes by 15-50 mV, often to a voltage at which action potential generation become inactivated after a few impulses, and the cell typically repolarizes more slowly (= 100 ms) (Matsumoto and Ajmone Marsan 1964a). The activation of excitatory receptors during the PDS activates voltage-sensitive channels and usually evokes action potentials that contribute to the generation and propagation of the METHODS activity (Gutnick et al. 1982; Johnston and Brown 198 1; All cultures were grown in medium containing synaptic Kandel and Spencer 196 1; Matsumoto and Ajmone Mar- blockers ( 11.3 mM Mg2+ and l-5 mM kynurenate), conditions 0022-3077/9 1 $1 SO Copyright 0 199 1 The American Physiological Society
761
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 16, 2019.
762
M.
M
that lead to epileptiform activity in mass cultures of neurons (Furshpan and Potter 1989) and microcultures (Segal and Furshpan 1990). The technique for constructing collagen-on-agarose microcultures has been described (Segal and Furshpan 1990). Briefly, drops of collagen solution were sprayed from an atomizer onto a dried film of agarose. The collagen dried to form islands of substrate that promoted cell attachment; the exposed agarose surface was resistant to cell attachment. An initial plating was made of cells from the neonatal rat hippocampal formation, dissociated using a technique based on that of Furshpan and Potter ( 1989). This resulted in a layer of glial cells restricted to the dots of applied collagen. A second plating of cells resulted in attachment of neurons as well as glial cells to the areas defined by the glial covering on the collagen dots. The neurons in each microculture remained isolated from neurons in other microcultures by the cell-free agarose surface separating the microcultures. Recordings were made 16-6 1 days after the second plating. The recording chamber was warmed to 30°C and the bath level was several millimeters deep. Fine-tipped microelectrodes were filled with potassium citrate (2 M); resistances were loo-250 MO before impalement and 130-800 MQ after impalement. Neurons were accepted for analysis if stable recordings were obtained with resting potentials more negative than -50 mV and action potentials could be elicited with depolarizing current; impalements were typically maintained for 2-4 h. Input resistances for 10 consecutive solitary neurons were 337 rt 133 (SD) Mfi; action potential heights measured on an oscilloscope were 62.1 + 4.6 (SD) mV; resting membrane potentials were 6 1.O + 7.2 (SD) mV. Signals were recorded with an Axoprobe amplifier and digitized and recorded with a Vetter 3000 PCM recording adaptor; figures were made with a Gould TA 2000 chart recorder. The microelectrodes used for intracellular injection experiments were similar, but filled with 1% potassium Lucifer yellow (Molecular Probes, Eugene, OR) in 50 mM potassium citrate to allow recording and injection with the same electrode. The dye was injected with 1-s pulses of 0.1-0.3 nA every 2 s for 5-10 min. Standard solutions were similar to those described previously (Furshpan and Potter 1989; Segal and Furshpan 1990). The growth medium, based on L- 15 CO, medium (Hawrot and Patterson 1979), contained synaptic blockers (11.3 mM Mg2+ and l-5 mM kynurenate). The standard perfusion solution, based on Hank’s balanced salt solution with 10% (vol/vol) Eagle’s Minimum Essential Medium (modified with Earle’s salts and 2 g/l sodium bicarbonate, without glutamine, from ICN Biochemicals, Costa Mesa, CA), contained 1 mM Mg2+ and no kynurenate. Several modifications to the microculture and recording techniques (Segal and Furshpan 1990) were employed. A more concentrated agarose solution was used (0.2% vs. 0.15%) to reduce beading of the agarose solution. Vitrogen 100 collagen was used (2.9 mg/ml; Collagen, Palo Alto, CA) for forming microcultures instead of human placental collagen. Growth medium contained more kynurenate (5 mM in 52 of 93 single-neuron microcultures vs. 1 mM in early studies), and growth medium applied beginning 1 day after the second plating of cells contained less rat serum (2 vs. 5%). Cultures were fed every week instead of every 2 wk. Perfusion medium contained no glutamine (in 85 of 93 single-neuron microcultures vs. 1.6 mM in early studies). The phenomena described in this paper were similar before and after these changes in technique, except that the yield of usable microcultures increased and sources of glutamate toxicity were reduced. I obtained 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and 2-OH-saclofen from Research Biochemicals (Natick, MA), 2amino-5-phosphonovalerate (D(-)APV) from Cambridge Research Biochemicals (Valley Stream, NY), and tetrodotoxin (TTX) and glycine from Sigma.
SEGAL RESULTS
Physical appearance of neurons in microcultures A phase-contrast micrograph of a microculture with a solitary neuron is shown in Fig. 1A. Cells from the hippocampal formation attached selectively to the Gland of collagen, avoiding the background layer of agarose gel. Neuronal processes in microcultures were clearly visible by the use of phase contrast optics in the first 3 wk after neuronal attachment, but later the processes became less distinct, until only neuronal cell bodies could be discerned on the layer of flat nonneuronal (presumably glial) cells. Electron micrographs of such microcultures showed that large numbers of processes were embedded in several layers of nonneuronal cells (A. Yee and M. Segal, unpublished observations). Neuronal processes were also visualized by intracellular injection of potassium Lucifer yellow. Fourteen such injections were performed on neurons in single-neuron microcultures. Using criteria to be described below, it was determined that seven of the neurons were excitatory, six were inhibitory, and one was of undetermined type. In all in-
FIG. 1. Microscopic appearance of a single-neuron microculture. A: phase-contrast micrograph of a collagen-on-agarosemicroculture. A solitary excitatory neuron was growing on g!ialcells for 24 days on an island of collagenon an agarosebackground. Bar represents 50 pm. B: fluorescence micrograph at the same magnification of the same neuron after injection with potassium Lucifer yellow.
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 16, 2019.
EPILEPTIFORM
ACTIVITY
IN HIPPOCAMPAL
jetted cells, neuronlike processes were evident within the microculture, and no neuronal processes were seen outside the microcultures except those processes already visible by phase contrast microscopy. Figure IB shows a fluorescence micrograph of the same (excitatory) neuron shown in part A of the figure. Characterization
of‘neuronal .
type
Intracellular recordings were made from 93 single-neuron microcultures. These solitary neurons were classified as excitatory or inhibitory on the basis of the autosynaptic (autaptic) action that followed action potentials. The terms excitatory and inhibitory are used here to describe the observed synaptic actions and do not imply that a neuron released only one transmitter or that a transmitter always had the same effects on different neurons (cf. Potter et al. 1981; Segal 1983). Recordings from solitary excitatory neurons showed characteristic electrophysiological changes as 5 mM kynurenate was washed out of the culture dish. Intracellular injection of current into solitary excitatory neurons in 5 mM kynurenate typically resulted in several action potentials (Fig. 2A,). Perfusion with a control solution lacking kynurenate resulted in the development of a plateau of depolarization after the initial action potential (A2) that became, on complete washout of kynurenate, a complex event with one or more action potentials followed by a wedge-shaped depolarization (A&. These events lasted 100 ms- 1 s, with shorter duration events when events were elic-
MICROCULTURES
763
ited more frequently: when the neuron of Fig. 2A was stimulated every 2 s instead of every 10 s, the duration of the PDS-like events fell from 400 to 200 ms. Similar duration changes were obtained in three of three other excitatory neurons tested. These depolarizing events will be referred to as PDS-like events because of their resemblance to PDSs in other systems (see DISCUSSION). Restoring kynurenate to the perfusion fluid abolished the complex and restored the isolated action potentials (not shown). Neurons were considered to be excitatory if PDS-like events occurred in standard perfusion solution and were blocked by kynurenate; 40 of the 93 single-neuron microcultures displayed PDS-like events, and all these events were blocked by kynurenate. The size of 38 of the 40 excitatory neurons was measured: diameter was taken as the geometric mean of the longest and shortest dimensions of the soma measured from a photograph. The mean soma diameter was 20.8 pm (SD 3.0 pm, SE 0.49 pm). Neurons were considered to be inhibitory if the hyperpolarization that followed the action potential in standard perfusion solution was reduced by bicuculline; 48 of the 93 single-neuron microcultures displayed such bicucullinesensitive hyperpolarizations. In these solitary inhibitory neurons in standard perfusion solution, the hyperpolarization usually had a early phase and a late phase (Fig. 2B,). Perfusion with a solution containing 50 PM bicuculline, which blocks y-aminobutyric acid-A (GABA,) receptors, abolished the early hyperpolarization but left intact the late hyperpolarization (Fig. 2&). Perfusion with a solution containing 1 mM phaclofen (which blocks GABA, receptors;
B1 Control
Bicuculline SO,IM
Phaclofen 1 mM Bicuculline 50 PM
FIG. 2. Pharmacological block of autaptic activity in single-neuron microcultures. Same scale applies to all records. A : an excitatory neuron alone. Current was injected intracellu larly in 90-ms pulses every 10 s as 5 mM kynurenate was washed out of the recording chamber by control perfusion solution containing no kynurenate. PDS-like events developed after washout of kynurenate. Indicated times are relative to the time control solution first reached the culture dish. Neuron had been in culture 40 days; resting potential was -62 mV. B: an inhibitory neuron alone. Action potentials occurred spontaneously. B,: there was a hyperpolarization after the action potential in standard perfusion solution. B,: early hyperpolarization was blocked by bicuculline, leaving a late hyperpolarization. B, : late hyperpolarization was not blocked bY phaclofen. Neuron had been in culture 59 days; resti ng potential was -70 mV.
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 16, 2019.
764
M. M. SEGAL
A From Kynurenate 5$it2? w
Kynurenate 5 mM
Expt,ded, in B
B
t-J
20 mV 1seC
FIG. 3. Washout of kynurenate from a microculture with a solitary excitatory neuron. Large depolarizing events (60 mV) and smaller (< 10 mV) depolarizing events arose on washout of kynurenate and disappeared with re-introduction of kynurenate. Record shows the first time kynurenate was washed out. Arrows show time of solution change. The single action potential on the rising phase of the large depolarizing event in B is not visible at this time resolution. Culture duration was 23 days. Resting potential at the start of the kynurenate washout was -65 mV.
Dutar and Nicoll 1988; Segal and Furshpan 1990) had no effect on the late hyperpolarization in five of five cases tested (Fig. 2B3). The mean soma diameter of the 48 inhibitory neurons was 16.2 pm (SD 2.2 pm, SE 0.32 pm), significantly differ........_......_.
B ..,..
.. . .. . .
ent from the excitatory neurons (P = 2 X 10-12, 2-tailed Student’s t test). Five of the 93 solitary neurons were insensitive to both kynurenate and bicuculline and were not placed in either the excitatory or the inhibitory category. Their mean soma
..I.:iKynurenate5mM:: .,......,.... .......... ... .
.,... ...
."'
.._.:..,,
_..,.,.,_... '.
‘._! .... ..:_.
,,
... .,
..:.. 2sec
c,
.:
... . ..,,,,’ .. . _: ..:..’ 20mV :’ F-J 2sec
FIG. 4. Lasting sustained depolarization in a solitary excitatory neuron. Washout of kynurenate led to a sudden transition to a sustained depolarization; neuron then continued to depolarize and there was a cessation of action potentials. As kynurenate was re-introduced, PDS-like events appeared transiently and neuron repolarized to a level more negative than original resting potential. Arrow shows the time when the kynurenate-containing solution began to enter the culture dish. Neuron had been in culture 34 days; resting potential was -68 mV.
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 16, 2019.
EPILEPTIFORM
ACTIVITY
IN HIPPOCAMPAL
MICROCULTURES
765
diameter was 15.3 pm (SD 1.2 pm, SE 0.53 pm), and they of the solitary excitatory neurons slowed the frequency of aborted were in culture for a mean 2 1.2 days (SD 3.3 days, SE 1.5 PDS-like events, and stronger hyperpolarization days), whereas the neurons with identified transmitter type the events. Hyperpolarizations after the PDS-like event were present were in culture for a mean 28.4 days (SD 8.4 days, SE 0.9 days); the difference was near significance at P = 0.06 using in 24 of the 40 solitary excitatory neurons (e.g., Figs. 3 and Student’s t test. 6A; cf. Matsumoto and Ajmone Marsan 1964a). In 3 1 of 33 solitary excitatory neurons with low noise recordings, small-amplitude depolarizing events could be Spontaneous events in excitatory neurons seen (Fig. 3). They had faster rising than falling phases, and Three types of spontaneous events were seen in excit- amplitude ranged up to 10 mV. The small depolarizing atory neurons on washout of kynurenate. These were PDS- events were absent in two excitatory neurons: both were like events, small depolarizing events, and sustained depo- neurons without spontaneous PDS-like events (e.g., Fig. larizations. In all excitatory neurons, PDS-like events oc- 2A) and the small depolarizing events were
in
I’..S..d.
Epileptifom Activity in Microcultures Containing One Excitatory Hippocampal Neuron MICHAEL
M. SEGAL
Department of Neurology, Massachusetts General Hospital Harvard Medical School, Boston, Massachusetts 02115 SUMMARY
AND
CONCLUSIONS
1. Paroxysmal depolarizing shifts (PDSs) occur during interictal epileptiform activity. Sustained depolarizations are characteristic of ictal activity, and events resembling PDSs also occur during the sustained depolarizations. To study these elements of epileptiform activity in a simpler context, I used the in vitro chronic-excitatory-block model of epilepsy of Furshpan and Potter and the microculture technique of Segal and Furshpan. 2. Intracellular recordings were made from 93 single-neuron microcultures. Forty of these solitary neurons were excitatory; their action potentials were replaced by PDS-like events or sustained depolarizations as kynurenate was removed from the perfusion solution. PDS-like events were similar to PDSs in intact cortex, mass cultures, and microcultures with more than one neuron. Small voltage fluctuations were also seen in solitary excitatory neurons in the absence of recorded action potentials. Sustained depolarizations developed in 5 of the 40 excitatory neurons. Forty-eight of the 93 solitary neurons were inhibitory, with bicuculline-sensitive hyperpolarizations after the action potential (ascribable to y-aminobutyric acid-A autapses). None of the solitary inhibitory neurons displayed sustained depolarizations. Five of the 93 neurons were insensitive to both kynurenate and bicuculline and were not placed in either the excitatory or the inhibitory category. 3. Both N-methyl-D-aspartate (NMDA) and non-NMDA glutamate receptors contributed to the PDS-like events and sustained depolarizations. Only a non-NMDA glutamate receptor component was evident for the small voltage fluctuations. 4. Intracellular recordings were also made from two-neuron microcultures, each containing one excitatory neuron and one inhibitory neuron. Sustained depolarizations developed in five microcultures, in each case only in the excitatory neuron. INTRODUCTION
and Department
of Neurobiology,
san 1964a,b; Rutecki et al. 1985). In contrast, sustained depolarizations are events that last seconds to minutes, and are associated with ictal activity (Hablitz 1987; Haglund and Schwartzkroin 1984; Kandel and Spencer 196 1; Matsumoto and Ajmone Marsan 1964b; Somjen et al. 1985). To study these elements of epileptiform activity, a potentially useful approach is to reduce the complexity of the system and examine the PDSs and sustained depolarizations in the simplest neuronal circuits in which they occur. A model of epilepsy has been described using mass cultures containing hundreds or thousands of neurons (Furshpan and Potter 1989). The cells are dissociated from the hippocampal formations of neonatal rats and are grown chronically in the presence of agents that suppress synaptic activity [elevated Mg2+ and kynurenate, a blocker of both N-methyl-D-aspartate (NMDA) and non-NMDA type glutamate receptors]. After several weeks, intense epileptiform activity can be recorded arising spontaneously when the blocking agents are removed. This activity displays many features of seizures experimentally induced in intact cortex: large PDS-like events; sudden increases in their frequency, often accompanied by sustained depolarizations; activation of inhibitory as well as excitatory mechanisms (Kandel and Spencer 196 1; Rutecki et al. 1985); “postictal” hyperpolarizations when the episodes are terminated (Kandel and Spencer 196 1); and extensive neuronal death when episodes are not terminated (Siesjo and Wieloch 1986). Microcultures containing one or several neurons have been used as a simple system to study the synaptic functions of sympathetic neurons (Furshpan et al. 1976, 1986). A microculture version of the mass culture epilepsy model has been described recently (Segal and Furshpan 1990). In this model, microcultures containing several neurons were grown chronically in the presence of the synaptic blockers kynurenate and elevated Mg2+. In microcultures containing 2- 18 neurons, on removal of the blockers, transitions to seizure-like activity and PDS-like events were observed that were similar to the epileptiform activity observed in mass cultures. I report here that both PDS-like events and sustained depolarizations can occur in microcultures that contain a single excitatory neuron. Some of these findings have been reported in preliminary form (Segal 1990).
Intracellular recordings from cortical neurons in intactanimal models of epilepsy have demonstrated that paroxysmal depolarizing shifts (PDSs) are characteristic of interictal epileptiform activity, and events resembling PDSs are also seen as part of the ictal activity. In vivo, PDSs are large, complex depolarizing events in which the cell rapidly depolarizes by 15-50 mV, often to a voltage at which action potential generation become inactivated after a few impulses, and the cell typically repolarizes more slowly (= 100 ms) (Matsumoto and Ajmone Marsan 1964a). The activation of excitatory receptors during the PDS activates voltage-sensitive channels and usually evokes action potentials that contribute to the generation and propagation of the METHODS activity (Gutnick et al. 1982; Johnston and Brown 198 1; All cultures were grown in medium containing synaptic Kandel and Spencer 196 1; Matsumoto and Ajmone Mar- blockers ( 11.3 mM Mg2+ and l-5 mM kynurenate), conditions 0022-3077/9 1 $1 SO Copyright 0 199 1 The American Physiological Society
761
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 16, 2019.
762
M.
M
that lead to epileptiform activity in mass cultures of neurons (Furshpan and Potter 1989) and microcultures (Segal and Furshpan 1990). The technique for constructing collagen-on-agarose microcultures has been described (Segal and Furshpan 1990). Briefly, drops of collagen solution were sprayed from an atomizer onto a dried film of agarose. The collagen dried to form islands of substrate that promoted cell attachment; the exposed agarose surface was resistant to cell attachment. An initial plating was made of cells from the neonatal rat hippocampal formation, dissociated using a technique based on that of Furshpan and Potter ( 1989). This resulted in a layer of glial cells restricted to the dots of applied collagen. A second plating of cells resulted in attachment of neurons as well as glial cells to the areas defined by the glial covering on the collagen dots. The neurons in each microculture remained isolated from neurons in other microcultures by the cell-free agarose surface separating the microcultures. Recordings were made 16-6 1 days after the second plating. The recording chamber was warmed to 30°C and the bath level was several millimeters deep. Fine-tipped microelectrodes were filled with potassium citrate (2 M); resistances were loo-250 MO before impalement and 130-800 MQ after impalement. Neurons were accepted for analysis if stable recordings were obtained with resting potentials more negative than -50 mV and action potentials could be elicited with depolarizing current; impalements were typically maintained for 2-4 h. Input resistances for 10 consecutive solitary neurons were 337 rt 133 (SD) Mfi; action potential heights measured on an oscilloscope were 62.1 + 4.6 (SD) mV; resting membrane potentials were 6 1.O + 7.2 (SD) mV. Signals were recorded with an Axoprobe amplifier and digitized and recorded with a Vetter 3000 PCM recording adaptor; figures were made with a Gould TA 2000 chart recorder. The microelectrodes used for intracellular injection experiments were similar, but filled with 1% potassium Lucifer yellow (Molecular Probes, Eugene, OR) in 50 mM potassium citrate to allow recording and injection with the same electrode. The dye was injected with 1-s pulses of 0.1-0.3 nA every 2 s for 5-10 min. Standard solutions were similar to those described previously (Furshpan and Potter 1989; Segal and Furshpan 1990). The growth medium, based on L- 15 CO, medium (Hawrot and Patterson 1979), contained synaptic blockers (11.3 mM Mg2+ and l-5 mM kynurenate). The standard perfusion solution, based on Hank’s balanced salt solution with 10% (vol/vol) Eagle’s Minimum Essential Medium (modified with Earle’s salts and 2 g/l sodium bicarbonate, without glutamine, from ICN Biochemicals, Costa Mesa, CA), contained 1 mM Mg2+ and no kynurenate. Several modifications to the microculture and recording techniques (Segal and Furshpan 1990) were employed. A more concentrated agarose solution was used (0.2% vs. 0.15%) to reduce beading of the agarose solution. Vitrogen 100 collagen was used (2.9 mg/ml; Collagen, Palo Alto, CA) for forming microcultures instead of human placental collagen. Growth medium contained more kynurenate (5 mM in 52 of 93 single-neuron microcultures vs. 1 mM in early studies), and growth medium applied beginning 1 day after the second plating of cells contained less rat serum (2 vs. 5%). Cultures were fed every week instead of every 2 wk. Perfusion medium contained no glutamine (in 85 of 93 single-neuron microcultures vs. 1.6 mM in early studies). The phenomena described in this paper were similar before and after these changes in technique, except that the yield of usable microcultures increased and sources of glutamate toxicity were reduced. I obtained 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and 2-OH-saclofen from Research Biochemicals (Natick, MA), 2amino-5-phosphonovalerate (D(-)APV) from Cambridge Research Biochemicals (Valley Stream, NY), and tetrodotoxin (TTX) and glycine from Sigma.
SEGAL RESULTS
Physical appearance of neurons in microcultures A phase-contrast micrograph of a microculture with a solitary neuron is shown in Fig. 1A. Cells from the hippocampal formation attached selectively to the Gland of collagen, avoiding the background layer of agarose gel. Neuronal processes in microcultures were clearly visible by the use of phase contrast optics in the first 3 wk after neuronal attachment, but later the processes became less distinct, until only neuronal cell bodies could be discerned on the layer of flat nonneuronal (presumably glial) cells. Electron micrographs of such microcultures showed that large numbers of processes were embedded in several layers of nonneuronal cells (A. Yee and M. Segal, unpublished observations). Neuronal processes were also visualized by intracellular injection of potassium Lucifer yellow. Fourteen such injections were performed on neurons in single-neuron microcultures. Using criteria to be described below, it was determined that seven of the neurons were excitatory, six were inhibitory, and one was of undetermined type. In all in-
FIG. 1. Microscopic appearance of a single-neuron microculture. A: phase-contrast micrograph of a collagen-on-agarosemicroculture. A solitary excitatory neuron was growing on g!ialcells for 24 days on an island of collagenon an agarosebackground. Bar represents 50 pm. B: fluorescence micrograph at the same magnification of the same neuron after injection with potassium Lucifer yellow.
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 16, 2019.
EPILEPTIFORM
ACTIVITY
IN HIPPOCAMPAL
jetted cells, neuronlike processes were evident within the microculture, and no neuronal processes were seen outside the microcultures except those processes already visible by phase contrast microscopy. Figure IB shows a fluorescence micrograph of the same (excitatory) neuron shown in part A of the figure. Characterization
of‘neuronal .
type
Intracellular recordings were made from 93 single-neuron microcultures. These solitary neurons were classified as excitatory or inhibitory on the basis of the autosynaptic (autaptic) action that followed action potentials. The terms excitatory and inhibitory are used here to describe the observed synaptic actions and do not imply that a neuron released only one transmitter or that a transmitter always had the same effects on different neurons (cf. Potter et al. 1981; Segal 1983). Recordings from solitary excitatory neurons showed characteristic electrophysiological changes as 5 mM kynurenate was washed out of the culture dish. Intracellular injection of current into solitary excitatory neurons in 5 mM kynurenate typically resulted in several action potentials (Fig. 2A,). Perfusion with a control solution lacking kynurenate resulted in the development of a plateau of depolarization after the initial action potential (A2) that became, on complete washout of kynurenate, a complex event with one or more action potentials followed by a wedge-shaped depolarization (A&. These events lasted 100 ms- 1 s, with shorter duration events when events were elic-
MICROCULTURES
763
ited more frequently: when the neuron of Fig. 2A was stimulated every 2 s instead of every 10 s, the duration of the PDS-like events fell from 400 to 200 ms. Similar duration changes were obtained in three of three other excitatory neurons tested. These depolarizing events will be referred to as PDS-like events because of their resemblance to PDSs in other systems (see DISCUSSION). Restoring kynurenate to the perfusion fluid abolished the complex and restored the isolated action potentials (not shown). Neurons were considered to be excitatory if PDS-like events occurred in standard perfusion solution and were blocked by kynurenate; 40 of the 93 single-neuron microcultures displayed PDS-like events, and all these events were blocked by kynurenate. The size of 38 of the 40 excitatory neurons was measured: diameter was taken as the geometric mean of the longest and shortest dimensions of the soma measured from a photograph. The mean soma diameter was 20.8 pm (SD 3.0 pm, SE 0.49 pm). Neurons were considered to be inhibitory if the hyperpolarization that followed the action potential in standard perfusion solution was reduced by bicuculline; 48 of the 93 single-neuron microcultures displayed such bicucullinesensitive hyperpolarizations. In these solitary inhibitory neurons in standard perfusion solution, the hyperpolarization usually had a early phase and a late phase (Fig. 2B,). Perfusion with a solution containing 50 PM bicuculline, which blocks y-aminobutyric acid-A (GABA,) receptors, abolished the early hyperpolarization but left intact the late hyperpolarization (Fig. 2&). Perfusion with a solution containing 1 mM phaclofen (which blocks GABA, receptors;
B1 Control
Bicuculline SO,IM
Phaclofen 1 mM Bicuculline 50 PM
FIG. 2. Pharmacological block of autaptic activity in single-neuron microcultures. Same scale applies to all records. A : an excitatory neuron alone. Current was injected intracellu larly in 90-ms pulses every 10 s as 5 mM kynurenate was washed out of the recording chamber by control perfusion solution containing no kynurenate. PDS-like events developed after washout of kynurenate. Indicated times are relative to the time control solution first reached the culture dish. Neuron had been in culture 40 days; resting potential was -62 mV. B: an inhibitory neuron alone. Action potentials occurred spontaneously. B,: there was a hyperpolarization after the action potential in standard perfusion solution. B,: early hyperpolarization was blocked by bicuculline, leaving a late hyperpolarization. B, : late hyperpolarization was not blocked bY phaclofen. Neuron had been in culture 59 days; resti ng potential was -70 mV.
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 16, 2019.
764
M. M. SEGAL
A From Kynurenate 5$it2? w
Kynurenate 5 mM
Expt,ded, in B
B
t-J
20 mV 1seC
FIG. 3. Washout of kynurenate from a microculture with a solitary excitatory neuron. Large depolarizing events (60 mV) and smaller (< 10 mV) depolarizing events arose on washout of kynurenate and disappeared with re-introduction of kynurenate. Record shows the first time kynurenate was washed out. Arrows show time of solution change. The single action potential on the rising phase of the large depolarizing event in B is not visible at this time resolution. Culture duration was 23 days. Resting potential at the start of the kynurenate washout was -65 mV.
Dutar and Nicoll 1988; Segal and Furshpan 1990) had no effect on the late hyperpolarization in five of five cases tested (Fig. 2B3). The mean soma diameter of the 48 inhibitory neurons was 16.2 pm (SD 2.2 pm, SE 0.32 pm), significantly differ........_......_.
B ..,..
.. . .. . .
ent from the excitatory neurons (P = 2 X 10-12, 2-tailed Student’s t test). Five of the 93 solitary neurons were insensitive to both kynurenate and bicuculline and were not placed in either the excitatory or the inhibitory category. Their mean soma
..I.:iKynurenate5mM:: .,......,.... .......... ... .
.,... ...
."'
.._.:..,,
_..,.,.,_... '.
‘._! .... ..:_.
,,
... .,
..:.. 2sec
c,
.:
... . ..,,,,’ .. . _: ..:..’ 20mV :’ F-J 2sec
FIG. 4. Lasting sustained depolarization in a solitary excitatory neuron. Washout of kynurenate led to a sudden transition to a sustained depolarization; neuron then continued to depolarize and there was a cessation of action potentials. As kynurenate was re-introduced, PDS-like events appeared transiently and neuron repolarized to a level more negative than original resting potential. Arrow shows the time when the kynurenate-containing solution began to enter the culture dish. Neuron had been in culture 34 days; resting potential was -68 mV.
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 16, 2019.
EPILEPTIFORM
ACTIVITY
IN HIPPOCAMPAL
MICROCULTURES
765
diameter was 15.3 pm (SD 1.2 pm, SE 0.53 pm), and they of the solitary excitatory neurons slowed the frequency of aborted were in culture for a mean 2 1.2 days (SD 3.3 days, SE 1.5 PDS-like events, and stronger hyperpolarization days), whereas the neurons with identified transmitter type the events. Hyperpolarizations after the PDS-like event were present were in culture for a mean 28.4 days (SD 8.4 days, SE 0.9 days); the difference was near significance at P = 0.06 using in 24 of the 40 solitary excitatory neurons (e.g., Figs. 3 and Student’s t test. 6A; cf. Matsumoto and Ajmone Marsan 1964a). In 3 1 of 33 solitary excitatory neurons with low noise recordings, small-amplitude depolarizing events could be Spontaneous events in excitatory neurons seen (Fig. 3). They had faster rising than falling phases, and Three types of spontaneous events were seen in excit- amplitude ranged up to 10 mV. The small depolarizing atory neurons on washout of kynurenate. These were PDS- events were absent in two excitatory neurons: both were like events, small depolarizing events, and sustained depo- neurons without spontaneous PDS-like events (e.g., Fig. larizations. In all excitatory neurons, PDS-like events oc- 2A) and the small depolarizing events were
