Effect of low glucose concentration on synaptic transmission in the rat hippocampal slice

Brain Research Bulletin, Vol. 21,

pp.

741-747.0

Pergamon Press

plc,

0361-9230/88 $3.00 + .OO

1988. Printed in the U.S.A.

Effect of Low Glucose Concentration on Synaptic Transmission in the Rat Hippocampal Slice’ P. FAN, P. A. O’REGAN Department

of Physiology and Biophysics,

AND J. C. SZERB2

Dalhousie

Received

University, Halifax, N.S.

Canada,

B3H 4H7

5 May 1988

FAN, P., P. A. O’REGAN AND J. C. SZERB. Ejfect oj’low glucose concentration on synaptic transmission in the rat hippocampal s/ice. BRAIN RES BULL 21(5) 741-747, 1988.-Severe hypoglycemia in vivo is known to slow down the EEG, then to produce complete electrical silence in the brain. To find out why low glucose concentrations reduce electrical activity, synaptic transmission from Schaffer collateraYcommissura1 fibers to CA1 pyramidal cells in the submerged rat hippocampal slice was investigated using extracellular recording techniques. Superfusion for 30 min with 1 mM glucose reversibly reduced population spike amplitude, without affecting the size of the presynaptic volley and the slope of the field EPSP. Lower glucose concentrations also affected the EPSP, although to a lesser extent than the population spike. Antidromic population spikes were not decreased by low glucose. Depolarization with 8-10 mM K+ reduced both presynaptic volley amplitude and EPSP, but enhanced the population spike, an effect clearly different from that of low glucose. The slope of the input/output curve between presynaptic volley and EPSP remained unaltered in 1 mM glucose but the slope between EPSP and population spike was reduced by about 50%. Results suggest that low glucose concentrations interrupt synaptic transmission by reducing, but not abolishing, the excitability of pyramidal cells. Hippocampal slice

Synaptic transmission

Low glucose

Synaptic inputiouput

Neuronal excitability

synaptic transmission in guinea-pig (2, 9, 15) and rat (19) hippocampal slices. Both presynaptic (15) and postsynaptic (2) factors were suggested to be responsible for this depression of synaptic transmission. The purpose of this study was to establish the site of the block in synaptic transmission by analyzing the effect of lowering the glucose content of the medium on the input/output relationship between presynaptic action potentials, excitatory postsynaptic potentials and population spikes in CA1 pyramidal cells of the rat hippocampus measured extracellularly.

INSULIN-INDUCED hypoglycemia has profound effects on the metabolism and electrical activity of the brain in vivo. Initially, as blood-sugar levels fall, EEG and behavioral changes appear without major alterations in the energy status of the brain (3,13). This is followed by a complete absence of electrical activity, accompanied by a large reduction in high-energy phosphates (7,11), a biphasic increase in extracellular potassium and decrease in extracellular calcium concentrations (1,23). Furthermore, there is an increase in aspartate and a reduction in glutamate content of the brain (6,14), and a large increase in extracellular aspartate and a smaller increase in extracellular glutamate concentrations (5,W. Some of the neurochemical changes induced by hypoglycemia in vivo have been observed in vitro: incubation in a low glucose medium increases the aspartate and reduces the glutamate content of brain slices (4,17). Furthermore, in low glucose solutions the depolarization-induced release of aspartate is greatly increased, while that of glutamate increases much less (21,22). Similarly to the in vivo situation, reducing the glucose content of the superfusion medium decreases

METHOD

Transverse, 0.4 mm thick, fully submerged hippocampal slices from male Sprague-Dawley rats were superfused at a rate of 1 mUmin at 32+1”C with Krebs solution saturated with 95% 0, and 5% CO,. The composition of the Krebs solution was (in mM): NaCl 120; CaC& 2.6; KC1 3.0; NaH,PO, 1.2; MgSO, 1.2; NaHCO, 25; glucose 5.0. After 90 min superfusion, the Schaffer collaterals near the CA2/CAl border were stimulated with single pulses at 0.1 Hz with

‘This work was supported by the Medical Research Council of Canada. P.F. was a recipient of a Dalhousie scholarship. 2Requests for reprints should be addressed to Dr. John C. Szerb.

741

University

Graduate

FAN. O’KEGAN 150 140

-

130

-

120

-

11a

-

1OO

-

90

-

80

-

70

-

60

1

-l_-.__l--_

----~

/

SO 40

-.-

150 140

B

A

130

-

120

-

110

-

.--.

C

loo90

-

80

-

70

-

60

-

50

-

150

A

140

0

4

6

12

16

B

i

20

24

28

32

36

40

44

f

46

TIME (min.)

52

56

60

C

64

66

72

76

c10 CL4 00

Y2

AND SZERB

LOW GLUCOSE

AND SYNAPTIC

ANTGMOMC

ORTtlODROMlC

GLUCOSE

:

_“_. :\

5mM

.. y/k*

/--\

‘, . :a-+

_.-_ \., ;‘-

i’ i:. . : _. :,:

( Control )

TRANSMISSION

.: .... ‘,

I’ ;: .. ’

‘ .

0.2 mM

..

:

_.-_ \ :2,2; ,.---__ .: .:

,_.“_~,__:~,

.. ::

( 25 min. )

.. ::

743

After the initial 90 min superfusion, the experiments consisted of A) a 30 min period during which the control Krebs solution containing 5 mM glucose and 3 mM K+ was perfused through the bath, B) a 30 min experimental period, during which the solution under investigation was perfused, and 3) a final 30 min period during which the controi solution was again applied. Results were accepted only when the responses returned to not less than 80% of their initial size after switching back to the control solution. With the exception of inpu~output curves, results were expressed as a percent of the average responses during the initial control period. Significance of the differences were calculated with the Student’s r-test. RESULTS

/‘*

__!:_‘\ ,

5 mM

( 27 min. after 0.2 mM )

i, ; . ‘-.__., ;, ::

*-_

- i p..__ . .*.___,_ __: ,: , :: .. . ‘. ;: :’

r

CA~BflA~N

0.1 mV L5ms

FIG. 2. Effect of 0.2 mM glucose (middle traces) on population spikes evoked orthodromically (stimulation in the stratum radiatum) and antidromically (stimulation in the alveus) by paired stimuli, 10 msec apart. Note the marked potentiation of the population spike due to the second orthodromic but not due to the second antidromic stimulation.

monopolar tungsten electrodes with pulses of 0.0.5-O. 1 msec duration and 50-150 pAMP intensity to obtain o~hodromic excitation of the CA1 pyr~idal neurons. Field potentials from CA1 stratum radiatum and stratum pyramidale were recorded with glass microelectrodes filled with Krebs solution (resistance of 2-4 Mohm). The intensity of stimulation was chosen to give 2/3-314of the maximal responses. For comparing the effect of lowering the glucose concentration on orthodromic and antidromic excitation of CA1 pyramidal cells, paired stimuli, 10 msec apart, were applied alternately to either the alveus or the Schtier collaterals and the population spike in the pyramidal layer was measured. Paired stimuli clearly distinguished between orthodromic and antidromic excitation since only orthodromic stimulation resulted in facilitation (Fig. 2). Recorded potentials were digitized and stored on diskettes for analysis. The program analyzed for the size of the presynaptic volley and the maximal rate of depolarization (mV/msec) during the field-EPSP (fEPSP) in the stratum radiatum and for the size of the population spike in the stratum pyramidale.

Ejyect of’ Low Glucose on Orthodromic Excitation Superfusion with a solution containing 0.2-1.0 mM glucose caused a rapid, reversible, concentration-dependent depression in the size of the population spike produced by Schaffer collatera~commissur~ fiber stimulation (Fig. 1, top). A smail but signi~cant decrease in population spike amplitude was produced by 1 mM glucose and this depression remained nearly constant for 20 min, allowing an input/output analysis with several intensities of stimulation (see later). In 0.2 and 0.5 mM glucose the maximal slope of the fEPSP was reduced to a lesser extent than the population spike and 1 mM glucose did not affect the fEPSP (Fig. 1, middle). Even the lowest concentration of glucose failed to decrease the presynaptic volley (Fig. 1, bottom). Ejyect

of Low

Glucose on Antidromic

Excitation

To see whether the reduction of population spike is the result of a decreased excitability of pyramidal neurons in low glucose, the effect of 0.2 mM glucose on the ~pulation spike evoked by paired o~hodromic and antidromic stimulation was compared. The second of the o~hodromic~ly evoked population spikes, which was considerably potentiated by the conditioning stimulus, disappeared in 0.2 mM glucose, while both of the antidromically evoked population spikes, which were about equal, remained unchanged in low glucose (Fig. 2). Effect oj‘ Elevnted [K+10 on Orthodromic

Transmission

As indicated in the introduction, severe hypoglycemia in vivo is accompanied by an elevation of [K+], in the CNS, hence by depolarization. There was a possibility that in brain slices too, a reduction in the supply of glucose would lead to a similar increase in [K+],, due to a reduction in the activity of the Na/K pump. Altematively, a reduced supply of glucose could lead directly to depolarization, for instance, by an increase in sodium conductance. To see if depolarization per se could account for the observed changes in synaptic transmission in low glucose, the effect of increasing the con-

FACING PAGE FIG. 1. Effect of super-fusion with reduced glucose concentrations on the population spike amplitude (top), fEPSP maximal slope (middle) and presynaptic voliey amplitude (bottom). During periods marked A and C, 5.0 mM glucose was present in all experiments. In the period marked by B the glucose concent~tion was changed to 0.2 mM (squares); to 0.5 mM (crosses); to 1.O mM (diamonds) or left unchanged at 5.0 mM (triangles). Results were no~aIi~d by making the average of the observation during the initial supe~usion with 5.0 mM glucose in each experiment equai 100%. Average of 6 experiments shown. Asterisks denote significant Q~0.05) differences from 100%.

744

FAN, CI’REGAN 260

:

i

i :

A

AND SZERH

C

l

I :

L-

160

100 80 60 40 20 170

my

_~

:

I

160

6

A

150 140

I

60 50 40 30 20

-

10 0

11

1’1%

4

8

12

I ( 1’1) 16

20

17 24

28

32

36

40 TfMf

44

48

52

56

60

64

68

72

76

80

84

88

42

(MIN.)

FIG. 3. Effect of 8 mM (tap) and IO m&l (bottamt K’ on the amplitude of the presynaptic volley (squares), maximai slope of fEPSP fcrosses), and ~pu~a~~on spike amplitude ~diarno~ds~~During periods marked A and C, 3.0 mM K” was present, while the concentration of potassium was changed during period B. Average of 6 experiments. Results were normalized as in Fig. I.

74.5

GLUCOSE AND SYNAPTIC T~NSMISSION

PPeY#l ley MY

FIG. 4. Example of the effect of 1 mM glucose on the inpu~output relationship between presynaptic volley and fEPSP maximal slope (top) and fEPSP maximal siope and ~p~ation spike amplitude (bottom) calculated From data obtained from the same experiment. Rectangles and broken lines during the initial suffusion with 5.0 mM glucose; stars and solid lines during superfusion with 1.0 mM glucose; triangles and dots during the final superfusion with 5.0 mM glucose. Lines were fitted by the least square method. Exposure to 1 mM glucose for 30 min.

centration of potassium to 8 or 10 mM in the presence of S mM glucose was investigated. ~h~~n~ the [I(+], to 8 mM had little effect on the size of the presynaptic volley, but reduced the maximal slope of the fEPSP by about 60%. Simultaneously, the size of the population spikes increased

about 2.5 times (Fig. 3, top). Increasing [K+l, to IO mI’v% almost completely elim~ated the presyuaptic volley and the fJZPSP towards the end of the 30 min superfusion. In contrast, it depressed the population spike by only 40% and even potentiated it at the beginning of the exposure to 10 mM K”

746

FAN, O’REGAN TABLE

I

EFFECT OF 1 mM GLUCOSE ON THE SLOPES OF THE INPUT/OUTPUT LINES

Glucose

Concentration

Presynaptic Volley/fEPSP

tEPSP/Population Spike

5 mM (initial) 1 mM 5 mM (final)

Slopes were normalized by making the slopes in the initial 5 mM glucose solution equal IO@%.Averages i s.e.m. of six observations. *Highly s~gni~cantly &
and at the beginning of the washout period (Fig. 3, bottom).

The observation that reducing the glucose content of the medium depressed the population spike more than the WPSP, suggested that it affected the excitability of the soma more than transmission from the synaptic terminals to the dendrites. To confirm this, the inputioutput relationships between presynapt~c volley/~PSP and ~PSPlpopulation spike were established by stimulation with at least five different intensities before, during and after 30 min superfusion with I mM glucose. Less than 1 mM glucose depressed population spike amplitude to such an extent that even the highest intensity of stimulation failed to increase it. In 1 mM glucose the input~output relationship between presynaptic volley and FEPSP was not affected (Fig. 4, top, Table l), but the slope of the input/output curve between fE?PSP and population spike measured simultaneously was reversibly depressed by more than 50% (Fig. 4, bottom, Table I). Changing the concentration of glucose from 5 mM to 2 mM (two experiments) or to 10 mM (3 experiments) had no effect on either of these input/output curves (not shown).

DISCUSSION

Present observations confirm previous findings (2, 9, 19) on the reversible inhibition of synaptic transmission in the hippocampus in vitro produced by low glucose concentrations. However, there are also a few differences between previous and present observations. The use in this study of fully submerged slices resulted in faster changes in transmission upon altering the concentration of glucose than observed in interface slices (19). On the other hand, signals obtained from submerged slices in this study were smaller than those recorded from slices at the air/medium interface. In the present experiments these smaller signals did not decrease and even increased during the 90 min experimental period (Fig. l), indicating that a deterioration in the condition of the slices was not the reason for the smaller size of the signals. Rather a shunting of the signals by the medium is likely to account for the small size of the potentiaIs measured. Several of the possible mechanisms leading to a depression in synaptic t~nsmission in low glucose can be excluded by the present findings. First of all, it failed to affect axOnal

AND S%ERB

conduction in presynaptic fibers or in pyramidal cell :txons which are responsible for the generation of antidromi~ poputation spikes. Neither is a reduction in the release from synaptic terminals of the excitatory transmitter. probably glutamate and/or aspartate (8, 12, 16), likely to be responsible for a decrease in synaptic transmission in low giucosc. This is suggested by the observation that a moderate reduction in glucose content to 1 mM, does not change the fjrst p~~stsynaptic event. the fEPSP and the pl.esyrl~~pti~ voiIeyitEPSP input~output slope. Similarly. no change in granule cell fEPSP was found when the glucose content 01 the medium was lowered from 10 to 5 mM, yet the population spike was reduced (9). Furthermore, it is known that the release of aspartate and glutamate evoked by electrical-field stimulation from hippo~ampal slices is not decreased in low glucose, but increased (21.22). Observations on the effects of elevated potassium concentrations clearly excluded depolarization of either the presynaptic or postsynaptic neurons as the reason for transmission failure in a low glucose environment: milder depolarization with 8 mM K’, while reducing somewhat the maximal WPSP slope, greatly increased the population spike ~~mpiitude, white IO mM K’ decreased the amplitude of the presynaptic volley and nearly suppressed fEPSP, while decreasing population spike amplitude only moderately. These changes, which are probably the result of multiple events, such as a decreased firing threshold of the postsynaptic neurons and a depolarization block of the presynaptic axons, are clearly different from those observed in a low glucose environment. Furthermore. the depressions in synaptic transmission described here occur in a glucose concentration which maintains the ATP and phosphocreatine content of hippocampal slices (10). ‘Therefore, changes in synaptic transmission described here are unlikely to be the consequence ofan insufficient energy supply. Rather, they appear to be a specific effect of glucose deprivation. The highest concentration of glucose (I mM) found to impair synaptic transmission reduced only the slope of the fEPSP/population spike input/output curve, without affecting that of the presynaptic volley/fEPSP. Only a more drastic reduction in glucose concentration to 0.5 or 0.2 mM glucose, which completely eliminated orthodromically evoked population spikes, reduced the fE2PSP. This reduction was much iess and occurred more slowly than the decrease in population spike. However. even the lowest concentration of glucase used, 0.2 mM, had no effect on antidromically evoked population spikes. This suggests that the suppression of orthodromically produced population spike is not the result of an absolute inability of pyramidal cell somata to fire, but only to an increase in their firing threshold. Such an increase would affect more the triggering of action potentials by a graded depolarization, such as an EPSP, than by an all-ornothing depolarization induced by antidromic stimulation. The main finding of this study, namely that a reduced supply of glucose depresses synaptic transmission by an increase in the firing threshold of the postsynaptic neuron, is at variance with the conclusions of Miyakawa f’f ttl. (1%. namely that the block is presynaptic. They based their conclusions on extra- and intracellular studies on the effect of total glucose deprivation in the guinea-pig hippocampus which resulted in a block in transmission much before the excitability of the postsynaptic neuron to antidromic or direct stimulation was suppressed. However. the severe and progressive depression in synaptic transmission due to a total absence of glucose precluded an analysis of the rela-

LOW GLUCOSE

AND SYNAPTIC

TRANSMISSION

tionship between the EPSP and action potential threshold. In contrast, the conclusions of Bachelard er al. (2,9) on the postsynaptic site of action of low glucose analogues in depressing the synaptic activation of dentate granule cells of the guinea-pig are similar to that reached in the present

study. Furthermore, a recent study (20) showed that glucose-free media causes a hyperpolarization of guinea-pig CA3 pyramidal neurons due to an increase in potassium conductance, an effect which could explain the increased firing threshold of these neurons in low glucose.

REFERENCES 1. Astrup, J.; Norberg, K. Potassium activity in cerebral cortex in rats during progressive severe hypoglycemia. Brain Res. 103:41&423; 1976. 2. Bachelard, H. S.; Cox, D. W. G.; Drawer, J. Sensitivity of guinea-pig hippocampal granule cell field potential to hexoses in vitro: an effect on cell excitability? J. Physiol. (Lend.) 352:91102; 1984. 3. Behar, K. L.; den Hollander, J. A.; Petroff, 0. A. C.; Hetherington, H. P.; Prichard, J. W.; Shuhnan, R. G. Effect of hy~~ycemic encephalopa~hy upon amino acids, high-energy phosphates, and pHi in the rat brain in vivo: detection by sequential ‘H and $‘P NMR spectroscopy. J. Neurochem. 44:1045-1055; 1985. 4. Benjamin, A. M.; Quastel, J. H. Location of amino acids in brain slices from rat. Tetrodotoxin-sensitive release of amino acids. Biochem. J. 128631-646; 1972. 5. Butcher, S. P.; Hagberg, H.; Sandberg, M.; Hamberger, A. Extracellular purine catabolite and amino acid levels in the rat striatum during severe hypoglycemia: effect of 2amino-5phosphonovalerate. Neurochem. Int. 11:95-99; 1987. 6. Butterworth, R. F.; Merkel, P. L.; Landreville, F. Regional amino acid distribution in relation to function in insulin hypoglycemia. J. Neurochem. 38: 1488-1489; 1982. 7. Chapman, A. G.; Westerberg, E.; Siesjo, B. K. The metabolism of purine and py~midine nucleotides in rat cortex during insulin-induced hy~glycemia and recovery. J. Neurochem. 36:179-189; 1981. 8. Corradetti, R.; Moneti, G.; Moroni, F.; Pepeu, G.; Wieraszko, A. Electrical stimulation of the stratum radiatum increases the release and neosynthesis of aspartate, glutamate and y-aminobutyric acid in rat hippocampal slices. J. Neurochem. 4l:f518-1525; 1983. 9. Cox, D. W. G.; Bachelard, H. S. Attenuation of evoked field potentials from dentate granule cells by glucose, pyruvate f malate, and sodium fluoride. Brain Res. 239:527-534; 1982. 10. Cox, D. W. G.; Morris, P. G.; Feeney, J.; Bachelard, S. 3*P NMR studies on cerebral energy metabolism under conditions of hypoglycemia in vitro. Biochem. J. 212:365-370; 1983. 11. Gore&J. M.; Dolkart, P. H.; Ferendelh, J. A. Regional levels of glucose, amino acids, high energy phosphates and cyclic nucleotides in the central nervous system during hypoglycemic stupor and behavioral recovery. J. Neurochem. 27:1043-1049: 1976. 12. Hablitz, J. J.; Langmoen, I. A. N-methyl-d-aspartate receptor antagonists reduce synaptic excitation in the hippocampus. J. Neurosci. 6: 102-106: 1986.

13. Ikeda, M.; Yoshida, S.; Busto, R.; Santiso, M.; Martinez, E.; Ginsberg, M. D. Cerebral phosphoinositide and energy metabolism during and after insutin-induced hy~~ycemia. J. Neurothem. 49:100-106; 1987. 14. Lewis, L. D.; Ljungren, B.; Norberg, K.; Siesjo, B. K. Changes in carbohydrate substrate, amino acids and ammonia in the brain during insulin-induced hypoglycemia. J. Neurochem. 23:659-671; 1974. 15. Miyakawa, H.; Kaneko, K.; Kato, H. Effects of glucose deprivation on electrical activities of guinea-pig hip~~~l neurons studies in vitro. Yamagata Med. J. 3:13-26; 1985. 16. Ottersen, 0. P.; Storm-Mathisen, J. Different neuronal iocalization of aspartate-like and glutamate-like immunoreactivities in the hippocampus of rat, guinea-pig and Senegalese baboon (Papio Papio), with a note on the distribution of y-aminobutyrate. Neuroscience 16:.589-606; 1985. 17. Potashner, S. J. Spontaneous and electricaJly evoked release, from slices of guinea-pig cortex, of endogenous amino acids labelled via metabolism of d[Uy4CJglucose. J. Neurochem. 31:177-186; 1978. 18. Sandberg, M.; Butcher, S. P.; Hagberg, H. Extracellular overflow of neuroactive amino acids during severe insulin-induced hypoglycemia: In vivo dialysis of the rat hippocampus. .I. Neurochem. 47:178-184; 1986. 19. Schurr, A.; West, C. A.; Tseng, M. T.; Reid, K. H.; Rigor, B. M. Role of glucose in rn~nt~ni~ synaptic activity in the rat hippocampal slice preparation. In: Scherr, A.; Teyler, T. J.; Tseng, M. T., eds. Brain slices: Fundamentals, applications and implications. Basel: Karger; 1987:3%44. 20. Spuler, A.; Enders, W.; Grafe, P. Glucose depletion hyperpolarizes guinea pig hippocampal neurons by an increse in potassium conductance. Exp. Neurol. 100:248-252; 1988. 21. Szerb, J. C. Changes in the relative amounts of aspartate and glutamate released and retained in hippocampal slices during stimulation. J. Neurochem. 50321%224; 1988. 22. Szerb, J. C.; O’Regan, P. A. Reversible shifts in the Ca2+dependent release of aspartate and glutamate from hippocampal slices with changing glucose concentrations. Synapse 1:265272; 1987. 23. Wieloch, T. J.; Haris, R. J.; Symon, L.; Siesjo, B. K. Influence of severe hy~glycemia on brain extracellular calcium and potassium activities, energy and phospholipid metabolism, J. Neurochem. 43:160-168: 1984.