Distinct pharmacological and functional properties of NMDA receptors in mouse cortical astrocytes

Abstract

Background and purpose: Astrocytes of the mouse neocortex express functional NMDA receptors, which are not blocked by Mg(2+) ions. However, the pharmacological profile of glial NMDA receptors and their subunit composition is far from complete.

Experimental approach: We tested the sensitivity of NMDA receptor-mediated currents to the novel GluN2C/D subunit-selective antagonist UBP141 in mouse cortical astrocytes and neurons. We also examined the effect of memantine, an antagonist that has substantially different affinities for GluN2A/B and GluN2C/d-containing receptors in physiological concentrations of extracellular Mg(2+).

Key results: UBP141 had a strong inhibitory action on NMDA receptor-mediated transmembrane currents in the cortical layer II/III astrocytes with an IC(50) of 2.29 µM and a modest inhibitory action on NMDA-responses in the pyramidal neurons with IC(50) of 19.8 µM. Astroglial and neuronal NMDA receptors exhibited different sensitivities to memantine with IC(50) values of 2.19 and 10.8 µM, respectively. Consistent with pharmacological differences between astroglial and neuronal NMDA receptors, NMDA receptors in astrocytes showed lower Ca(2+) permeability than neuronal receptors with P(Ca) /P(Na) ratio of 3.4.

Conclusions and implications: The biophysical and pharmacological properties of the astrocytic NMDA receptors strongly suggest that they have a tri-heteromeric structure composed of GluN1, GluN2C/D and GluN3 subunits. The substantial difference between astroglial and neuronal NMDA receptors in their sensitivity to UBP141 and memantine may enable selective modulation of astrocytic signalling that could be very helpful for elucidating the mechanisms of neuron-glia communications. Our results may also provide the basis for the development of novel therapeutic agents specifically targeting glial signalling.

Figure 1

The GluN2C/D subunit-selective antagonist UBP141 differentially suppresses NMDA receptors in the astrocytes and neurons. (A) Representative traces illustrate the current activated by 50 µM NMDA in the acutely isolated cortical layer II/III astrocyte before and after application of 3 and 10 µM UBP141. (B) Effect of the same concentrations of UBP141 on current activated by 50 µM NMDA in a pyramidal neuron acutely isolated from the same cortical slice. Note the much weaker inhibitory effect of UBP141 in the neuron. (C) The mean concentration-dependent effect of UBP141 in seven astrocytes (IC₅₀, 2.29 ± 0.37 µM; Hill coefficient, 1.19 ± 0.14) and six neurons (IC₅₀, 19.8 ± 3.4 µM; Hill coefficient, 1.08 ± 0.15). The amplitude of the NMDA response was normalized to control. Error bars represent SD. Holding membrane potential was −80 mV in astrocytes and −40 mV in neurons.

Figure 4

Voltage-dependence of the currents mediated by NMDA receptors in astrocytes and neurons. The upper panels show currents induced by rapid application of NMDA (50 µM, 1 s) recorded at the different holding potentials in a cortical astrocyte (A) and pyramidal neuron (B) in 2 and 20 mM extracellular Ca²⁺. The lower panels show the I–V curves constructed from 11 (astrocytes) and 6 (neurons) independent experiments. The amplitudes of the responses to NMDA were normalized to the value measured at −40 mV; data are presented as mean ± SD. Solid lines show the results of a best polynomial fit (least squares routine), intersection with zero current axis gives the following values of reversal potential in 2 and 20 mM Ca²⁺_OUT: 0.92 mV and 7.5 mV for astrocytes and 2.75 and 14.1 mV for neurons. The permeability ratio P_Ca/P_Na calculated in the framework of extended Goldman-Hodgkin–Katz equation is 3.4 for astrocytes and 7.5 for neurons.

Figure 2

Memantine differentially inhibits NMDA receptors in astrocytes and neurons. (A) Representative traces illustrate the NMDA-activated current recorded in the acutely isolated cortical layer II/III astrocyte before and after application of 1 and 10 µM of memantine. (B) Effect of the same concentrations of memantine on an NMDA-activated current in a pyramidal neuron acutely isolated from the same cortical slice. Note the much weaker inhibitory effect of memantine in the neuron. (C) The mean concentration-dependent effect of memantine in seven astrocytes at −80 mV (IC₅₀, 2.19 ± 0.18 µM; Hill coefficient, 1.01 ± 0.07), in five astrocytes at −40 mV (IC₅₀, 4.24 ± 0.47 µM; Hill coefficient, 1.03 ± 0.11) and in six neurons at −40 mV (IC₅₀, 10.8 ± 0.55 µM; Hill coefficient, 1.04 ± 0.05). Error bars represent SD.

Figure 3

Effect of UBP141 and memantine on the NMDA-receptor mediated synaptic currents in astrocytes and neurons in situ. Transmembrane currents evoked in the astrocytes (glial synaptic currents; GSCs) and pyramidal neurons (epscs) of layer II/III of mouse neocortex in situ by stimulation of neuronal afferents were recorded in presence of 100 µM picrotoxin, 50 µM CNQX and 30 µM PPADS. (A) Left panel shows the changes in the amplitude of astrocytic GSCs and neuronal epscs following the bath application of 3 µM UBP141 and 10 µM ifenprodil. Data are presented as mean ± SD for 9 cells. (B) Changes in the astrocytic GSCs and neuronal epscs after application of 1 and 10 µM memantine (mean ± SD for eight cells). Data shown in (A) and (B) were measured in the constant presence of glutamate transporters antagonists TFB-TBOA, 1 µM and DL-TBOA, 30 µM. Each point in the time graphs shows the average amplitude (relative to control) of 5 sequential currents. Illustrative examples of GSCs and epscs (average of five traces) recorded as indicated, are shown in the right panels. Note the different inhibitory effects produced by UBP141, ifenprodil and memantine in the astrocytes and neurons. (C) Left panel shows the changes in the amplitude of astrocytic GSCs measured immediately after the establishment of whole-cell recording using intracellular solution supplemented with 10 µM MK-801 (mean ± SD for six cells). D-AP5, UBP141 and a combination of TFB-TBOA (1 µM) and DL-TBOA (30 µM) were applied to cortical slices as indicated. Note that UPB141 did not decrease the amplitude of the GSCs, when astrocytic NMDA receptors were inhibited by intracellular MK-801, confirming that its action shown in (A) was due to inhibition of glial NMDA receptors.

Figure 5

NMDA receptor-mediated Ca²⁺ signalling in cortical astrocytes. (A) An acutely isolated astrocyte was loaded with Fluo-3 via a somatic patch pipette. Fluorescent images were recorded simultaneously with transmembrane currents evoked by application of 20 µM NMDA and 100 µM glutamate in control and after consecutive application of 30 µM D-AP5. Representative images (left) and glial currents (right, upper row) were recorded before (rest) and after application of agonist. Ca²⁺ transients (right, lower row) represent the ΔF/F₀ ratio averaged over the cell soma; scale bar is 10 µm. Holding potential is −80 mV. Transmembrane current activated by glutamate in control is mediated by NMDA and AMPA receptors and glutamate transporters. The NMDA-component was eliminated by D-AP5. The Ca²⁺ transient activated by glutamate in the control was mediated by metabotropic glutamate receptors, NMDA receptors and possibly by Ca²⁺-permeable AMPA receptors. The Ca²⁺ transient activated by glutamate in the presence of D-AP5 lacks the NMDA-receptor-mediated component. (B) Pooled data (mean ± SD for five astrocytes) of peak Ca²⁺ transients; the difference is statistically significant with *P < 0.01 and **P < 0.005 (one-way anova). Note the significant contribution made by the NMDA receptors to the glutamatergic Ca²⁺ signal in the astrocytes.

Figure 6

Effect of glycine and D-serine on astroglial and neuronal NMDA receptors. (A, B) Representative responses activated in acutely-isolated cortical astrocyte and neuron by consecutive application of 50 µM NMDA (no glycine added), 10 µM glycine, 10 µM D-serine and 20 µM NMDA and D-serine together. (C) Pooled data (mean ± SD for 10 astrocytes and 7 neurons) of amplitude of current activated by glycine, D-serine and NMDA and D-serine together; amplitudes were normalized to the amplitude of the NMDA-activated current. Recordings were made at a holding potential of −80 mV in the astrocytes and −40 mV in the neurons. Note the weak agonist action of glycine and D-serine in the astrocytes.

Figure 7

Effect of NMDA receptor antagonists on glycine and D-serine-activated currents in astrocytes. (A) Representative responses activated in acutely-isolated cortical astrocytes by consecutive application of 50 µM NMDA (no glycine added) and 10 µM glycine in control and the presence of 10 µM MK801. NMDA and glycine were applied in turn with a 2 min interval. The gradual decline of the astrocytic response is indicative of a use-dependent block; block by MK801 established after the fourth to fifth round of application. Note the synchronous decrease in the amplitude of both NMDA and glycine-activated currents. (B) Representative responses activated in acutely-isolated cortical astrocytes by consecutive applications of 50 µM NMDA (no glycine added) and 10 µM D-serine in control and after the application of 10 µM memantine. NMDA and D-serine were applied in turn with a 2 min interval. (C) Pooled data (mean ± SD for five astrocytes) on the inhibitory effect of MK801, memantine and UBP141 on astrocytic responses. Although glycine-activated currents were less sensitive to the antagonists than the NMDA-activated currents, the difference was not statistically significant. All recordings were made at a holding potential of −80 mV in the presence of picrotoxin and strychnine.