Protection by imidazol(ine) drugs and agmatine of glutamate-induced neurotoxicity in cultured cerebellar granule cells through blockade of NMDA receptor

Abstract

This study was designed to assess the potential neuroprotective effect of several imidazol(ine) drugs and agmatine on glutamate-induced necrosis and on apoptosis induced by low extracellular K+ in cultured cerebellar granule cells. Exposure (30 min) of energy deprived cells to L-glutamate (1-100 microM) caused a concentration-dependent neurotoxicity, as determined 24 h later by a decrease in the ability of the cells to metabolize 3-(4,5-dimethythiazol-2-yl)-2,5-diphenyltetrazoliumbromide (MTT) into a reduced formazan product. L-glutamate-induced neurotoxicity (EC50=5 microM) was blocked by the specific NMDA receptor antagonist MK-801 (dizocilpine). Imidazol(ine) drugs and agmatine fully prevented neurotoxicity induced by 20 microM (EC100) L-glutamate with the rank order (EC50 in microM): antazoline (13)>cirazoline (44)>LSL 61122 [2-styryl-2-imidazoline] (54)>LSL 60101 [2-(2-benzofuranyl) imidazole] (75)>idazoxan (90)>LSL 60129 [2-(1,4-benzodioxan-6-yl)-4,5-dihydroimidazole](101)>RX82 1002 (2-methoxy idazoxan) (106)>agmatine (196). No neuroprotective effect of these drugs was observed in a model of apoptotic neuronal cell death (reduction of extracellular K+) which does not involve stimulation of NMDA receptors. Imidazol(ine) drugs and agmatine fully inhibited [3H]-(+)-MK-801 binding to the phencyclidine site of NMDA receptors in rat brain. The profile of drug potency protecting against L-glutamate neurotoxicity correlated well (r=0.90) with the potency of the same compounds competing against [3H]-(+)-MK-801 binding. In HEK-293 cells transfected to express the NR1-1a and NR2C subunits of the NMDA receptor, antazoline and agmatine produced a voltage- and concentration-dependent block of glutamate-induced currents. Analysis of the voltage dependence of the block was consistent with the presence of a binding site for antazoline located within the NMDA channel pore with an IC50 of 10-12 microM at 0 mV. It is concluded that imidazol(ine) drugs and agmatine are neuroprotective against glutamate-induced necrotic neuronal cell death in vitro and that this effect is mediated through NMDA receptor blockade by interacting with a site located within the NMDA channel pore.

Figure 1

(a) Neurotoxic effect of L-glutamate on cerebellar granule cells as measured by the concentration-dependent inhibition in the formation of reduced MTT relative to unexposed control cells. Neurons were depleted of energy resources by 40 min preincubation in a medium without magnesium and glucose and then incubated for 30 min in the presence of different concentrations of L-glutamate. Neurotoxicity was determined 24 h later as described in Methods. Data shown are mean±s.e.mean of eight wells. (b) Prevention by MK-801 (dizocilpine) of the neurotoxic effect of an EC₁₀₀ of L-glutamate in cerebellar granule cells. Data shown are mean±s.e.mean of eight wells. ^*P<0.001 as compared with cells exposed to L-glutamate (Student's t-test).

Figure 2

Neuroprotective effects of (a) antazoline, cirazoline, RX821002 and idazoxan and (b) LSL 61122, LSL 60129, LSL 60101 and agmatine against the neurotoxic effect of 20 μM L-glutamate in cerebellar granule cells, as measured by the concentration-dependent increase in the formation of reduced MTT relative to L-glutamate exposed cells. Neurons were depleted of energy resources by 40 min preincubation in a medium without magnesium and glucose and then incubated for 30 min in the presence of L-glutamate. Imidazol(ine) drugs were added during the last 15 min of the preincubation time and maintained throughout the L-glutamate exposure. Neuroprotection was determined 24 h later as described in Methods. Points represent mean±s.e.mean of eight wells. See Table 1 for EC₅₀ values.

Figure 3

Effect of cirazoline, LSL 60101 and LSL 61122 on neurotoxicity induced by switching from serum-free high [K⁺]_out (25 mM) to serum-free low [K⁺]_out (5 mM) medium in cerebellar granule cells. Cells were maintained in serum-free Eagle's Basal Medium containing either 25 or 5 mM K⁺ for 24 h in the presence (100 μM) or absence (control) of drugs where indicated. Neurotoxicity was measured using the MTT reduction (mitochondrial activity) assay. Data shown are mean±s.e.mean of absorbance (optical density, O.D.) units from eight wells. ^*P<0.05 as compared with cells maintained in high [K⁺]_out (25 mM) medium (Student's t-test). No significant differences were found between control cells and cells exposed to drugs and maintained in low [K⁺]_out (5 mM) medium.

Figure 4

Inhibition of binding of [³H]-(+)-MK-801 by (a) antazoline, cirazoline, idazoxan and RX821002 and by (b) LSL 61122, LSL 60129, LSL 60101 and agmatine in the rat cerebral cortex. Membranes were incubated at 23°C for 45 min with [³H]-(+)-MK-801 (4×10⁻⁹ M) in the absence or presence of different concentrations of the competing drugs. Data shown are mean±s.e.mean of 2–4 independent experiments performed in triplicate and expressed as a percentage of total control binding (about 12,000 d.p.m.). See Table 1 for K_i values.

Figure 5

Correlation between the potency (expressed as pK_i=−log K_i) of several imidazol(ine) drugs and agmatine inhibiting the binding of [³H]-(+)-MK-801 to NMDA receptors in the rat brain and the potency (expressed as pEC₅₀=−log EC₅₀) of the same compounds protecting against L-glutamate neurotoxicity in cerebellar granule cells (data taken from Table 1). The data were best described by the equation y=0.75x+1.06 (r=0.90; P<0.01). The identification of drugs is as follows: (1) antazoline, (2) cirazoline, (3) LSL 61122, (4) LSL 60101, (5) idazoxan, (6) LSL 60129, (7) RX821002 and (8) agmatine.

Figure 6

NMDA receptor channel blockade by antazoline in HEK-293 cells transfected to express the NR1-1a and NR2C subunits of the NMDA receptor. (a) Response to 200 μM glutamate in the absence (control) or presence of the indicated concentrations of antazoline. The response returned to control levels upon removal of antazoline (not shown). The current traces were interrupted by a voltage ramp to +70 mV (shaded bar). On depolarization the channel was partially relieved of antazoline blockade. On returning to the −70 mV holding voltage the channel was blocked again, causing the observed relaxations of the traces on the right. (b) Ramp I–V curves in the absence (control) or the presence of the indicated concentrations of antazoline. Solid lines represent fits to the Boltzman equation. (c) Extent of current reduction at different voltages in the presence of the indicated concentrations of antazoline. This relationship was generated from the fitted relation in (b) (see Methods). The vertical lines represent the s.e.mean of at least four independent determinations. (d) The logarithm of the IC₅₀ values for the inhibition of glutamate-induced currents are plotted as a function of voltage. The relation presented two linear portions that were fitted to the Woodhull model; the best fit parameters for IC₅₀ at 0 mV and zδ are indicated on the figure.

Figure 7

NMDA receptor channel blockade by agmatine in HEK-293 cells transfected to express the NR1-1a and NR2C subunits of the NMDA receptor. (a) Responses to 200 μM glutamate without (control) or with 200 μM agmatine. The response returned to control levels upon removal of agmatine (not shown). The current trace is interrupted by a voltage ramp to +70 mV (shaded bar). (b) Ramp I–V curves obtained in the absence (control) or the presence of 200 μM agmatine in the same cell as in (a). Solid lines represent fits to the Boltzman equation. These data were obtained from the same cell as in Figure 6a. (c) Extent of current reduction at different voltages in the presence of the indicated concentrations of agmatine. This relationship was generated from the fitted relation in (b) (see Methods). The vertical lines represent the s.e.mean of at least four independent determinations .