Control of Glutamate Clearance and Synaptic Efficacy by Glial Coverage of Neurons

See allHide authors and affiliations

Science  04 May 2001:
Vol. 292, Issue 5518, pp. 923-926
DOI: 10.1126/science.1059162


Analysis of excitatory synaptic transmission in the rat hypothalamic supraoptic nucleus revealed that glutamate clearance and, as a consequence, glutamate concentration and diffusion in the extracellular space, is associated with the degree of astrocytic coverage of its neurons. Reduction in glutamate clearance, whether induced pharmacologically or associated with a relative decrease of glial coverage in the vicinity of synapses, affected transmitter release through modulation of presynaptic metabotropic glutamate receptors. Astrocytic wrapping of neurons, therefore, contributes to the regulation of synaptic efficacy in the central nervous system.

Astrocytes contribute to the regulation of synaptic transmission by controlling glutamate diffusion and concentration in the extracellular space (1–3). Changes in the glial coverage of neurons in the vicinity of synapses may thus alter glutamate clearance and synaptic transmission (4). To examine the effect of changes in glial coverage on synaptic transmission, we recorded from the magnocellular nuclei of the hypothalamus, which undergo a well-documented anatomical neuroglial plasticity in response to intense stimulation, like lactation (5, 6). This results in a decreased coverage of neurons by astrocytic processes and a relative absence of these processes in the vicinity of the synapses. The changes are reversed with cessation of stimulation in postlactating animals.

To investigate such neuroglial interactions, we first examined the functional consequences of a glutamate clearance deficiency on excitatory synaptic transmission. We inhibited glutamate transporters in the rat supraoptic nucleus (SON) with either dihydrokainate (DHK)—a specific inhibitor of GLT-1 (7, 8), a glutamate transporter exclusively expressed in astrocytes (9–11)—orl-trans-pyrrolidine-2,4-dicarboxylic acid (PDC), a broad-spectrum glutamate transporter blocker (8). In virgin rats, where glial coverage is extensive (5,6), whole-cell voltage-clamp recordings (12) of SON neurons revealed that evoked excitatory postsynaptic currents (EPSCs) were inhibited reversibly (Fig. 1, A and B) by 100 μM DHK (46.6 ± 8.0% of control values, n = 9) and 300 μM PDC (51.2 ± 6.8%, n = 12). That PDC and DHK depressed evoked EPSCs to the same extent suggests that, for the most part, their effect was mediated by inhibition of GLT-1 transporters, which are present only on astrocytes in this and other brain areas (11). The depression in EPSC amplitude induced by glutamate transporter antagonists was associated with an increase in the paired-pulse facilitation (PPF) ratio (13), from 1.2 ± 0.1 to 1.6 ± 0.2 (40-ms interval; P < 0.05,n = 6). Analysis of miniature EPSCs (mEPSCs) (Fig. 1, C to E) revealed that PDC decreased the frequency (−34.0 ± 6.1%;P < 0.05, n = 6), but not the size (−1.6 ± 4.3%; P > 0.05), of these events (Fig. 1F). This indicates that the decrease in excitatory synaptic transmission associated with blockade of glutamate transporters had a presynaptic origin.

Figure 1

Glutamate transporter blockade induces presynaptic inhibition of EPSCs. (A) Sample traces of evoked EPSCs obtained in SON neurons in the presence and absence of PDC (upper panel) or DHK (lower panel). Traces are averages of 10 consecutive responses obtained from virgin, lactating (L8), and postlactating (post) rats. For illustrative purposes, EPSCs are scaled to responses obtained in virgin rats under control conditions. Stimulus artifacts are removed (arrow). (B) Summary histogram of the inhibitory actions of PDC and DHK on evoked EPSCs in the three groups of animals. In each group, the reduction obtained with PDC was not statistically different from that obtained with DHK (P> 0.05); the number of cells is indicated in parentheses. (C) Example of a cell showing PDC-induced reduction in mEPSC activity. The corresponding cumulative amplitude distributions (D) obtained in the presence and absence of PDC were not statistically different (P > 0.05), whereas the cumulative event interval distribution (E) was significantly shifted to the right with PDC (P > 0.05), which corresponds to a reduction in mEPSC frequency. (F) Summary of the effects of PDC on mEPSC size (q) and frequency (Hz) observed in six different cells.

These effects could result from a local build-up of glutamate that would lead to activation of presynaptic glutamate receptors controlling neurotransmitter release (14, 15). Group III metabotropic glutamate receptors (mGluRs) are known to induce a presynaptic inhibition in the SON (16). We thus investigated the action of PDC in the presence ofl-2-amino-4-phosphonobutyric acid (L-AP4) and 2-amino-2-methyl-4-phosphonobutanoic acid (MAP4), an agonist and an antagonist of group III mGluRs, respectively (Fig. 2, A and C). Whereas L-AP4 (200 μM) impaired evoked EPSCs (19.6 ± 2.8%; P < 0.05,n = 5), subsequent addition of PDC had no further effect on EPSC amplitude (22.8 ± 3.0%; P > 0.05, n = 5). In contrast, 250 μM MAP4 increased EPSC amplitude (130.0 ± 10.0%; P < 0.05,n = 9), confirming the existence of a tonic activation of these receptors (16). MAP4 prevented the effects of PDC (131.2 ± 14.5%; P > 0.05,n = 5). These findings indicate that blockade of glutamate transporters in the SON modifies the concentration and/or diffusion of glutamate in the extracellular space, leading to activation of presynaptic group III mGluRs and inhibition of glutamate release.

Figure 2

Glutamate transporter blockade activates presynaptic group III mGluRs. (A) Sample traces of evoked EPSCs obtained in virgin rats in the absence and presence of group III mGluR agonist (L-AP4, upper panel) and antagonist (MAP4, lower panel). Subsequent application of PDC had no significant effect on EPSCs amplitude under both conditions. (B) Sample traces of evoked EPSCs obtained in the presence and absence of MAP4 in lactating (L8) and postlactating (post) rats. (C) Summary graph illustrating the changes in EPSC amplitude measured in virgin, lactating, and postlactating animals under different conditions; the number of cells is indicated in parentheses.

To investigate whether similar changes in excitatory synaptic transmission occurred under physiological conditions that modify glial coverage of SON neurons, we examined hypothalamic slices obtained from lactating and postlactating animals. Glutamate transporter blockade was less effective in reducing evoked EPSCs in lactating rats (Fig. 1, A and B). Application of either DHK or PDC depressed EPSCs to 78.2 ± 10.3% (n = 8) and 75.0 ± 5.1% (n = 7) of control values, respectively. EPSC inhibition elicited by glutamate transporter blockade in postlactating animals was similar to that recorded in virgin rats (50.4 ± 7.4% in DHK, n = 8, and 46.6 ± 8.0% in PDC,n = 7) (Fig. 1, A and B). It appears, therefore, that the inhibitory action of glutamate transporter blockers on EPSCs is correlated with the degree of glial coverage of neurons. To determine whether reduction in astrocytic coverage affects presynaptic group III mGluRs, we compared the effect of MAP4 on EPSC amplitude in the three groups of animals. MAP4 induced a larger increase (P < 0.05) of evoked EPSC amplitude (163.3 ± 12.1%, n= 7) (Fig. 2, B and C) than observed in virgin or postlactating (129.1 ± 11.3%, n = 4) rats. This is in agreement with an increase in tonic activation of mGluRs by ambient glutamate in lactating rats.

We next examined whether glutamate clearance deficiency also affects glutamate concentration and/or time course in the synaptic cleft (17, 18), using γ-d-glutamylglycine (γ-DGG), a low-affinity, competitive AMPA receptor antagonist whose effect is sensitive to the concentration and/or time course of glutamate in the synaptic cleft (19). γ-DGG (0.5 mM) significantly (P < 0.05) reduced mEPSC amplitude, to 49.1 ± 3.3% (n = 8, virgin rats) and to 50.5 ± 2.9% (n = 5, postlactating rats) of control values (Fig. 3). Addition of PDC in the presence of γ-DGG restored mEPSC amplitude to 67.0 ± 5.1% and to 63.5 ± 6.2% in virgin (P < 0.05) and postlactating rats (P < 0.05), respectively. In contrast, inhibition of mEPSCs (46.5 ± 5.8%, n = 6, virgin rats; 49.4 ± 6.1%, n = 5, postlactating rats) by CNQX (6-cyano-7-nitroquinoxaline-2,3-dione, 500 nM) was not affected by PDC (44.0 ± 4.2% and 50.8 ± 3.9%, respectively;P > 0.05), as expected from a high-affinity, slowly dissociating, competitive AMPA receptor antagonist (19). In slices from lactating rats, the γ-DGG–induced inhibition was smaller than in virgin and postlactating rats (67 ± 4%;P < 0.05, n = 11) and was unaffected by PDC (64 ± 4%; P > 0.05). Moreover, in lactating rats, the reduction of mEPSC amplitude by CNQX was similar to that measured in virgin and postlactating rats (45 ± 4%,n = 6) and was unaffected by PDC (46 ± 4%;P > 0.05). These data strongly suggest that the concentration and/or time course of glutamate in the synaptic cleft is indeed greater in lactating than in virgin and postlactating rats.

Figure 3

Glial coverage of SON neurons affects the relative glutamate concentration and/or time course in the synaptic cleft. Summary graph illustrating the inhibition of mEPSCs observed under different conditions in virgin, lactating, and postlactating animals. Data are expressed as the ratio (I drug/I control) of the average mEPSC amplitude measured in the presence of a given drug (I drug) and the average mEPSC amplitude obtained under control conditions (I control). Averages of mEPSCs obtained under the different conditions are shown above the histograms (the number of cells is indicated in parentheses).

Taken together, our findings indicate that reduced astrocytic coverage of SON neurons in lactating rats leads to a glutamate clearance deficiency, resulting in enhanced glutamate concentration in the extracellular space, increased activation of presynaptic mGluRs, and thus, presumably, to a lower probability (Pr) of glutamate release. In agreement with this hypothesis, the PPF ratio measured for short intervals (40 to 80 ms) was greater (P < 0.05) in lactating rats (Fig. 4A). Furthermore, in lactating animals, MAP4 restored the PPF ratio to values measured in virgin and postlactating animals, whereas in the latter, PDC increased PPF to values similar to those measured in lactating rats (Fig. 4B). Although we cannot exclude alterations in composition and/or number of postsynaptic receptors in lactating animals, as has been shown for GABA-A receptors (20), such modifications cannot account for the changes in Pr revealed by PPF measurements. These findings demonstrate, therefore, that astroglial wrapping of SON neurons, by controlling glutamate clearance by way of the DHK-sensitive GLT-1 transporter, plays a significant role in regulating the efficacy of glutamatergic neurotransmission.

Figure 4

Differences in transmitter release associated with neuroglial remodeling. (A) Superimposed traces obtained from virgin, lactating, and postlactating animals where paired-pulse stimulation was given at various intervals (left); the summary graph (right) illustrates the paired-pulse ratio measured in the three groups of animals. (B) Responses elicited 40 ms apart in the presence and absence of PDC for virgin and postlactating rats, and in the presence and absence of MAP4 for lactating animals (left). The corresponding summary graph illustrates the PPF ratio (right); the number of cells is indicated in parentheses.

Presynaptic inhibition can be overcome, to some extent, by frequency-dependent facilitation of transmitter release. Thus, it could serve as a high-pass filter reducing nonspecific background activity and increasing signal-to-noise ratio for information transmitted through high-frequency bursts of afferent synaptic potentials. Such a process would be favored by a relatively low degree of glial coverage of neurons, allowing an enhanced level of activation of presynaptic glutamate receptors. In hypothalamic magnocellular nuclei in particular, this mechanism may contribute to the synaptically driven high-frequency bursts of action potentials (40 to 80 Hz) characterizing oxytocinergic neurons during milk ejection (21,22). In addition to modulating transmitter release (14, 15), variations in glutamate concentration associated with neuroglial remodeling could also affect the amplitude and/or time course of EPSCs (17, 18), glutamate spillover (14, 23), and glutamate receptors on glial cells (24). Such modulations of neuronal and glial activity are probably relevant in other brain functions in which such plasticity has been implicated, including learning, control of development, and estrous cycle (25–27).

  • * To whom correspondence should be addressed: E-mail: stephane.oliet{at}


Stay Connected to Science

Navigate This Article