Brain Research, 528 {1990) 212-222 Elsevier

212 BRES 15817

Extracellular ions during veratridine-induced neurotoxicity in hippocampal slices: neuroprotective effects of flunarizine and tetrodotoxin David Ashton, Roland Willems, Roger Marrannes and Paul A.J. Janssen Department of Neuropsychopharmacology, Janssen Research Foundation, Beerse (Belgium) (Accepted 6 March 1990)

Key words: Sodium channel; Tetrodotoxin; Veratridine; Spreading depression; Ischemia; Excitotoxicity; Ion-selective electrode; Calcium; Sodium; Potassium; Hippocampus; Guinea pig; Ca2÷ antagonist

Veratridine, by blocking Na ÷ channel inactivation and shifting activation to more negative membrane potentials5, causes Na+-influx and a persistent tendency for depolarization. Veratridine is neurotoxic to cultured neurones43, and this neurotoxicity can be blocked by the class IV calcium antagonist, flunarizine37. We were interested to know whether similar effects could be found in a functional differentiated tissue containing adult neurones and glial cells. We examined this in hippocampal slices using extracellular potential recordings and ion-selective microelectrodes sensitive to [Na+]o, [Ca2+]o and [K÷]o. Veratridine blocked synaptic transmission in CA1, and induced several episodes of spreading depression (SD). This was followed by a long-lasting increase in [K+]o and a continuous decrease in [Ca+]o. Following veratridine exposure to hypoxia only revealed a small negative DC shift and small shifts in extracellular ions; indicating that the cells had lost the ability to maintain ion homeostasis before the hypoxia, and that veratridine had been neurotoxic. In hippocampal slices obtained from guinea pigs which had been pretreated with 40 mg/kg x 2 flunarizine orally the time before the first SD induced by veratridine was doubled. Although the ion shifts during the first SD were similar to controls, flunarizine reduced the time of recovery of [Ca2+]o, [K+]o and DC potential. The increase in [K+]o baseline and the massive decrease in [Ca2+]o baseline seen following the SDs in the solvent group were smaller in the flunarizine-treated slices. During the subsequent hypoxic period the negative DC shift was 8x larger in the flunarizine group, and the shifts in [K+]o, [Na+]o and [Ca2+]o were bigger. Tetrodotoxin also delayed the first SD during veratridine and increased the size of the DC shift during the subsequent hypoxic period. Both flunarizine and tetrodotoxin therefore protected adult brain tissue containing glia from the neurotoxicity of veratridine. These findings suggest that persistent Na+-influx and the consequent CaZ+-influx produce neurotoxicity, and that the ability to attenuate this neurotoxicity may be important in the mechanism of action of cerebroprotective drugs from different pharmacological classes. INTRODUCTION In recent years there has been an accumulation of both in vitro and in vivo d a t a suggesting a role for neurotransmission in ischemic neuronal d a m a g e 7'45. Whilst loss of cellular Ca2+-homeostasis is still thought to be the final harbinger of neuronal death 44, it is less clear how intracellular Ca a+ becomes raised to neurotoxic levels. N e u r o t r a n s m i s s i o n can increase intracellular Ca 2+ in at least 3 ways. Firstly m e m b r a n e depolarization by excita t o r y transmitters can o p e n the different types of voltagesensitive calcium channels 33. Secondly, r e c e p t o r - o p e r ated calcium channels, such as those activated by N M D A - r e c e p t o r agonists, can substantially increase intracellular Ca z+ (see ref. 22). Finally, transmitters coupled to a G - p r o t e i n second messenger system can cause Ca2+-release from endoplasmic reticulum via inositol1,4,5-triphosphate or a n o t h e r inositolphosphate derivative 46. In addition to the well-known excitotoxicity of

glutamate and other excitatory aminoacids 35, other neurotransmitters m a y also be important. It has recently been suggested that d o p a m i n e contributes to ischemic injury in the striatum and m a y even m o d u l a t e the effects of glutamate 9'1°. If neurotransmission is a key factor in ischemic cell death then application of the Na+-channel blocker tetrodotoxin ( T F X ) should a m e l i o r a t e damage. Prenen et al. 39 have shown that local application of T F X affords protection in the Levine m o d e l of hypoxic-ischemic encephalopathy. Veratridine, which holds o p e n the TTXsensitive Na+-channel 5, also p r o d u c e s neurotoxicity in neuronal cultures 37'43 and cardiac myocytes 51. This neurotoxicity, in serum-free foetal rat h i p p o c a m p a l cultures and in cardiomyocytes, is blocked by T T X and the Class IV Ca2+-antagonist, flunarizine 37,51. A l t h o u g h , flunarizine has been shown to be protective in m a n y in vivo models of ischemia 53, b o t h flunarizine and Ca2÷-antagonists of o t h e r classes are only p o o r l y active against

Correspondence: D. Ashton, Department of Neuropsychopharmacoiogy, Janssen Research Foundation, B-2340 Beerse, Belgium. 0006-8993/90/$03.50 t~ 1990 Elsevier Science Publishers B.V. (Biomedical Division)

213 neurotoxicity

induced

in c u l t u r e d

foetal neurones

150

by

3

g l u t a m a t e o r k a i n a t e 15'37. W e w e r e t h e r e f o r e i n t e r e s t e d to examine

whether

100

flunarizine would protect hippo50

c a m p a l slices c o n t a i n i n g f u n c t i o n a l a n d m a t u r e n e u r o n e s a n d glia f r o m a v e r a t r i d i n e a g g r e s s i o n . F u r t h e r we u s e d i o n - s e l e c t i v e m i c r o e l e c t r o d e s in a n a t t e m p t t o get c l o s e r to t h e m e c h a n i s m o f a c t i o n o f v e r a t r i d i n e a n d f l u n a r i z i n e .

0

mV -50

-100

MATERIALS AND METHODS -150

Preparation Hippocampal slices of 400 p m thickness were prepared from 300-350 g male guinea pigs exactly as described previously 4. Slices were transferred to an interface, laminar-flow chamber at 35 °C and superfused with artificial cerebrospinal fluid (ACSF) at a rate of 3 ml/min. Normal ACSF was composed of (mM): NaC1 134, KCI 4, KH2PO 4 1.25, NaHCO 3 22, CaCI 2 2, MgSO 4 1.1. and D-glucose 10. The pH was 7.35 after saturation with 95% 02/5% CO 2. A stimulating electrode was positioned on the Schaffer collaterals and an extracellular electrode (filled with 2 M NaCI, 2-10 Mg2) or a double-barrelled liquid ion-sensitive microelectrode for Na ÷, K ÷ or Ca 2÷ was placed in the cell-body layer of CA1.

Ion-sensitive microelectrodes The ion-sensitive microelectrodes were fabricated as described in Marrannes et alY. The ion-exchangers were: K ÷, Coming 477317; Ca 2÷, Fluka cocktail 21048; and Na ÷, Fluka cocktail 71176. The K÷-sensitive microelectrodes were calibrated in solutions of 100, 10, 5.25 and 1 mM KCI with NaCI added to preserve the total concentration at 150 mM and thus keep ionic strength constant. Na+-sensitive microelectrodes were calibrated in solutions of 150, 100, 75, 50 and 25 mM NaC! with KCI added to maintain total concentration at 150 raM. The responses of K ÷- or Na+-electrodes were fitted to the Nicholsky equation: EK+ = S log ([K +] + A K Na [Na+]) + Eo ENa+ = S log ([Na +] + ANa K [K+]) + Eo by least-squares fit yielding slope (S), selectivity coefficients (AK_Na and ANa_K) and E o. Ca2÷-sensitive microelectrodes were calibrated in solutions containing (mM): CaCI 2 10 + NaC1 135, CaCI 2 1 + NaCI 148.5, CaCI 2 0.1 + NaCI 150. Responses were fitted to the equation: Eca2+ = S log ([Ca 2+] + B) + E o by least-squares fit, where B is an operational constant 25. Tissue [CaZ+]o was expressed as that [Ca 2+] which would produce the same E in a similar calibration solution series. This [Ca2+]o is the fraction of the total dissolved Ca 2÷ which is not complexed by anions such as bicarbonate. All calibration solutions were warmed to 35 °C4z. The selectivity (mean + S.D.) and slope of the Na+-electrodes was 0.581 ( + 0.125) and 127.0 (+ 43.1) mV/decade, respectively. The mean selectivity and slope of the K÷-electrodes was 0.017 (+ 0.005) and 55.9 ( + 1.9) mV/decade, respectively. The mean value for B and slope of the Ca z+ electrodes was 0.022 ( _ 0.014) and 29.1 ( + 1.7) mV/decade, respectively.

2

-200

600

650

700

750

800

850

900

Time (see)

Fig. 1. The automatic analysis of spreading depressions (SD) calculated the following parameters: A, initial maximum (point 1); B, amplitude of SD (point 2-point 1); C, maximum recovery amplitude (point 3-point 1); D, maximum recovery time (time point 3-time point 1); E, 50% recovery time between 50% SD amplitude on descending arm of curve and same amplitude on ascending arm of curve; F, maximum slope of SD for every consecutive two data points between 1 and 2 the slope of the straight line connecting the two points was calculated. The maximum value was the maximum slope. Similar points were also calculated for ion shifts during SD using other custom MACROs.

Selected sections of Eio n were converted into ion-concentration using the values for slope, selectivity coefficient and E o obtained from the calibration by a specially written Excel Macro. These and the DC recordings were then plotted and analyzed visually using Statview 512 ÷, or automatically using another set of Excel Macro's. The points which were automatically calculated are shown in Fig. 1.

Experimental protocol Slices were allowed to recover for 1 h in the chamber before setting electrodes. Slices were stimulated at 0.05 Hz using a pulse of 100 ps at a voltage 50% above population spike threshold. After a 15-min stabilization period, slices were exposed to veratridine for 25 min. A wash period of 30 rain followed. Then the slices were made hypoxic for 5 min by changing the gas flowing over the upper surface of the slices from 95% 02/5% CO z to 95% N2/5% CO 2.

Drugs and solutions Veratridine (Sigma) was dissolved in absolute ethanol to give a stock solution, which was added to ACSF to give a concentration of 3.0 x 10-5 M. Final ethanol concentration was 0.004%. TTX (Janssen Biochimica) was dissolved in distilled water and used in ACSF at a concentration of 1.25 x 10 -6 M. Flunarizine (Janssen Pharmaceutica) was dissolved in 5% hydroxypropyl cyclodextrine diluted to give a concentration of 1 or 4 mg/ml and given orally (1 ml per 100 g b. wt.) to give a dose of 10 or 40 mg/kg, Guinea pigs received one dose of flunarizine at 16.00 h and the next dose 1 h before dissection the following day. 'Low' Ca 2÷ ACSF was normal ACSF minus the 2 mM CaCIz + 1 mM EGTA.

Data collection The DC potential from the extracellular electrode or from the reference barrel of the ion-sensitive electrodes was continually sampled at 20 Hz by a MacLab system run on a Macintosh II computer. The potential difference [Eio,] between the ion-sensitive barrel and that of the reference barrel was also sampled at 20 Hz continuously by the MacLab system. The monosynaptic evoked responses were averaged by a Gould digital oscilloscope and stored on a HP9816 computer for analysis.

Statistics The unpaired t-test, two-tailed was used to compare differences between groups. P ~ 0.05 was considered significant.

Flunarizine levels At the end of the experiment slices in the chamber were pooled and deep frozen for analysis of flunarizine levels55 and protein content.

214 RESULTS

Solvent

Flunarizine

O-

D C and e v o k e d potentials

1-

In normal A C S F containing 5.25 m M (K+]o, exposing hippocampal slices to veratridine led to a short-lived increase in size of the population spike and in excitatory postsynaptic potential (EPSP) slope. This was followed by a spreading depression (SD) like depolarization (Fig. 4) and by the loss of evoked responses. The first SD (SD 0 was significantly delayed by flunarizine 40 mg/kg (Table I). The size of SD1 was around 25 mV in solventand flunarizine-treated slices (Table I). The maximal slope, and 50% recovery times of the SD a were also similar in the 3 groups. H o w e v e r the time of maximal recovery and the size of the positive shift versus baseline was also significantly reduced by flunarizine 2 x 40 mg/kg (Table I). One slice from a flunarizine 2 x 40 mg/kg treated animal showed no SDs during the experiment. The SD 1 was often followed by a second SD ( S D 2 ) , its characteristics were similar to controls in slices made from animals treated with 2 x 10 mg/kg flunarizine. In slices from animals treated with flunarizine 2 x 40 mg/kg the size of the S O 2 w a s larger, and its time of occurrence was delayed (Table I). All other measured parameters were similar. O n e 2 x 40 mg/kg flunarizine-treated slice showed a third SD. Following the S D 2 the D C signal showed a slow negativity followed by a return to preveratridine levels. The slope of the slow negative shift was 0.46 + 0.11 mV/min in solvents and 0.18 _+ 0.04 mV/min in the flunarizine 2 × 40 mg/kg group (P = 0.01). Similar experiments were also conducted with extracellular K + levels of 3.25 and 7.25 raM. In the solvent groups increasing the K + level decreased the time of occurrence of the SD a (3.25 m M 628.8 s, 5.25 mM 411.1 s, 7.25 m M 408.6 s). O t h e r values were not significantly different. In 3.25 m M K + flunarizine 2 x 40 mg/kg only

: |- :

23>"

E

4-

"~

5-

~

7-

D

8 ~

~8

6-

10 11 12

•

13

Fig. 2. individual values (0), median (midline of box) and interquartial range (upper and lower lines of box) for maximal

negative extracellular DC potential shift in CA1 region during exposure of hippocampal slices to hypoxia (95% N2/5% CO2). Two-tailed unpaired t-test P < 0.0001 between slices from solventtreated animals vs slices from animals treated with flunarizine.

reduced the time of recovery from the SD 1 to 50% of control values (P = 0.01). In 7.25 m M K +, flunarizine 2 x 40 mg/kg had the same effects as in 5.25 m M K ÷, in addition it decreased the maximal slope of the negative D C shift during SD 1 from 100 + 16 mV/ms in solvents (n = 8) to 58 + 9 mV/ms (n = 8) (P = 0.04). Thus increasing [K+]o, i.e. reducing the m e m b r a n e potential, appeared to increase the effects of flunarizine. Following the 30-min wash period the hippocampal slices were made hypoxic by switching the humidified gas flowing over their upper surface from 95% 02/5% CO 2 to 95% N2/5% C O 2. W h e n exposed to hypoxia hippocampal slices which have not been treated with veratridine show a 20-25 m V negative D C shift (the anoxic depolarization) 2"13, which occurs as a result of a simultaneous rapid intracellular depolarization 4°. The size of the anoxic depolarization depends upon the m e m b r a n e

TABLE I Mean + S.E.M. values for parameters indicated in Fig. 1 for extracellular DC potential during 1st and 2nd spreading depression induced by veratridine in CA1 region of hippocampal slices obtained from guinea pigs treated orally with 2 x 10 mg/kg (n = 8), or 2 × 40 mg/kg (n = 8) flunarizine or its solvent (n = 10) Amplitude (mV)

Maximum recovery (mV)

24.58 + 2.83 24.83 + 2.18 24.72 + 3.81

4.29 + 0.74 5.266 + 0.98 2.293 + 0.49*

19.02 + 2.12 17.45 __+2.53 24.18 + 1.63"

1.83 + 0.51 2.83 + 0.59 2.152 + 0.71

Maximum recovery (s)

50% recovery (s)

SD time (s)

75.7 + 6.94 65.49 + 10.12 45.39 + 7.61"

11.65 + 0.87 15.49 + 4.51 8.295 + 1.46"

411.1 + 14.3 412.8 + 18.59 1159.4 + 309.8**

105.52 + 18.56 99.43 + 11.29 77.99 + 12.63

13.36 + 1.04 18.45 + 2.95 14.58 + 3.34

770.9 + 26.8 713.3 + 25.8 1272.2 + 107.7"*

SDI

Solvent 10 mg/kg 40 mg/kg SD2

Solvent 10 mg/kg 40 mg/kg

*P=< 0.05; **P =