Physiological activation of mGlu5 receptors supports the ion channel function of NMDA receptors in hippocampal LTD induction in vivo

Abstract

Synaptic long-term depression (LTD) is believed to underlie critical mnemonic processes in the adult hippocampus. The roles of the metabotropic and ionotropic actions of glutamate in the induction of synaptic LTD by electrical low-frequency stimulation (LFS) in the living adult animal is poorly understood. Here we examined the requirement for metabotropic glutamate (mGlu) and NMDA receptors in LTD induction in anaesthetized adult rats. LTD induction was primarily dependent on NMDA receptors and required the involvement of both the ion channel function and GluN2B subunit of the receptor. Endogenous mGlu5 receptor activation necessitated the local application of relatively high doses of either competitive or non-competitive NMDA receptor antagonists to block LTD induction. Moreover, boosting endogenous glutamate activation of mGlu5 receptors with a positive allosteric modulator lowered the threshold for NMDA receptor-dependent LTD induction by weak LFS. The present data provide support in the living animal that NMDA receptor-dependent LTD is boosted by endogenously released glutamate activation of mGlu5 receptors. Given the predominant perisynaptic location of mGlu5 receptors, the present findings emphasize the need to further evaluate the contribution and mechanisms of these receptors in NMDA receptor-dependent synaptic plasticity in the adult hippocampus in vivo.

Figure 1

Combined administration of antagonists of NMDA and mGlu5 receptors, when given systemically, is necessary to prevent the induction of hippocampal synaptic LTD by low-frequency conditioning stimulation in vivo. (a) The competitive NMDAR antagonist CPP (10 mg/kg, i.p.), injected alone 2.25 h prior to electrical LFS (900 high-intensity pulses at 1 Hz) (see also Supplementary Fig. S1), had no significant effect on LTD induction (CPP: 64.4 ± 10%, mean ± SEM fEPSP amplitude expressed as a % of the pre-LFS baseline, n = 10; controls: 66.2 ± 7.1%, n = 10; P < 0.05 both compared with respective pre-LFS baselines, P > 0.05 between groups; two-way ANOVA RM-Sidak). (b) Summary of the mean EPSP amplitude data in (a) before (pre) and 1 h after (post) application of LFS. (c) Injection of the mGlu5R negative allosteric modulator MTEP (3 or 6 mg/kg, i.p., hash) alone did not significantly alter the induction of LTD by LFS 1 h later (MTEP: 75.0 ± 3.8%, n = 9; controls: 64.3 ± 7.7%, n = 9; P < 0.05 both compared with baselines, P > 0.05 between groups, two-way ANOVA RM-Sidak). (d) Summary of the mean EPSP amplitude data in (c). (e) Combined administration of CPP (10 mg/kg, i.p., 2.25 h pre-LFS) with MTEP (3 mg/kg, i.p., 1 h pre-LFS, hash), blocked induction of LTD (CPP + MTEP: 96.7 ± 4.9%, n = 8, P > 0.05 compared with pre-LFS baseline; Veh: 60.6 ± 3.8%, n = 9, P < 0.05 compared with baseline and between groups; two-way ANOVA RM-Sidak). (f) Summary of the mean EPSP amplitude data in (e). *P < 0.05. Insets show typical fEPSP traces at the times indicated. Calibration bars: vertical, 1 mV; horizontal, 10 ms.

Figure 2

Group II mGluR or mGlu1R antagonists alone or in combination with NMDAR blockade fails to inhibit synaptic LTD in vivo. (a) The potent group II mGluR antagonist LY341495 (3 mg/kg, i.p.) did not significantly affect the induction of LTD (LY341495: 74.8 ± 5.4%, n = 11; Veh: 68.2 ± 2.9%, n = 10; P < 0.05 both compared with baselines, P > 0.05 between groups; two-way ANOVA RM-Sidak) (see also Supplementary Fig. S1). (b) Summary of the mean EPSP amplitude data in (a). (c) Intraperitoneal administration of CPP (10 mg/kg, 2.25 h pre-LFS) combined with LY341495 (3 mg/kg, 1 h pre LFS), had no effect on LTD (CPP + LY341495: 61.2 ± 3.7%, n = 5; Veh: 64.9 ± 2.4%, n = 7, P < 0.05 compared to respective baselines and P > 0.05 between groups; two-way ANOVA RM-Sidak). (d) Summary of the mean EPSP amplitude data in (c). (e) Neither injection of the selective mGlu1R antagonist JNJ16259685 (0.5 mg/kg, s.c., asterisk) alone, nor when combined with CPP (10 mg/kg, i.p., 2.25 h pre-LFS) significantly affected the induction of LTD (JNJ16259685: 63.1 ± 1.8%, n = 6; CPP + JNJ16259685: 66.6 ± 4.0%, n = 6; P < 0.05 both compared with baselines, P > 0.05 between groups; two-way ANOVA RM-Sidak). (f) Summary of the mean EPSP amplitude data in (e). *P < 0.05. Calibration bars: vertical, 1 mV; horizontal, 10 ms.

Figure 3

GluN2B-subunit-selective antagonist, systemically administered with an mGlu5R antagonist, prevents LFS induction of LTD. (a) The non-competitive GluN2B subtype selective NMDAR antagonist Ro 25-6981 (12 mg/kg, i.p., 1 h pre-LFS, hash) alone had no effect on LTD induction (Ro 25-6981: 63.1 ± 4.6%, n = 5; Veh: 64.3 ± 4.2%, n = 6; P < 0.05 both compared with baselines, P > 0.05 between groups; two-way ANOVA RM-Sidak). (b) Summary of the mean EPSP amplitude data in (a). (c) In contrast, a combination of Ro 25-6981 (12 mg/kg, i.p., first hash) with MTEP (3 mg/kg, second hash) strongly attenuated LTD (Ro 25-6981 + MTEP: 91.7 ± 2.1%, n = 6; Veh: 64.2 ± 1.4%, n = 7; P < 0.05 compared with respective baselines and between groups; two-way ANOVA RM-Sidak). (d) Summary of the mean EPSP amplitude data in (c). *P < 0.05. Calibration bars: vertical, 1 mV; horizontal, 10 ms.

Figure 4

Relatively high doses of locally-injected NMDAR antagonists prevent LTD induction by LFS in vivo. (a) Intracerebroventricular administration of 200 nmol of the competitive NMDAR antagonist D-AP5 (triangle), 10 min before LFS-900 attenuated LFS induced LTD (200 nmol D-AP5: 73.2 ± 3.9%, n = 5; P < 0.05 compared with Veh: 51.1 ± 3.0%, n = 7; and P < 0.05 compared with 200 nmol of the NMDAR low-affinity stereoisomer of D-AP5, L-AP5: 51.4 ± 2.1%, n = 5; one-way ANOVA-Sidak). The magnitude of LTD in animals injected with the lower dose 100 nmol D-AP5 was not significantly different from vehicle or 200 nmol D-AP5 (100 nmol D-AP5; 61.3 ± 4.5%, n = 5). All groups were significantly different from their respective baselines (paired t). (b) Summary of the mean EPSP amplitude data in (a). (c) Furthermore, i.c.v. application of the non-competitive Ro 25-6981 (2 nmol, triangle) strongly inhibited LTD (2 nmol Ro: 88.2 ± 3.9%, n = 5, P > 0.05 compared with baseline; Veh: 53.8 ± 3.1%, n = 8, P < 0.05 compared with baseline and between groups; two-way ANOVA RM-Sidak). (d) Summary of the mean EPSP data in (c). (e) I.c.v. injection of the lower dose Ro 25-6981 (0.5 nmol, triangle) alone did not affect LTD induction but completely prevented LTD when given in combination with MTEP (3 mg/kg, i.p., 1 h pre-LFS) (0.5 nmol Ro: 57.7 ± 6.6%, n = 6, P < 0.05 compared with baseline; 0.5 nmol Ro + MTEP: 86.4 ± 3.0%, n = 5, P > 0.05 compared with baseline, P < 0.05 between groups; two-way ANOVA RM-Sidak). (f) Summary of the mean EPSP amplitude data in (e). *P < 0.05. Calibration bars: vertical, 1 mV; horizontal, 10 ms.

Figure 5

Endogenous glutamate release activates mGlu5Rs to facilitate the induction of LTD in vivo. (a) The mGlu5R positive allosteric modulator VU 0360172 (15 mg/kg, s.c., asterisk) facilitated the induction of LTD by a peri-threshold electrical LFS-300 conditioning protocol (300 high-intensity pulses at 1 Hz) relative to vehicle controls (control: 90.1 ± 2.6%, n = 11; VU 0360172: 58.7 ± 4.0%, n = 8, P < 0.05 compared with baselines and between groups; one-way ANOVA-Sidak followed by paired t). The facilitation of LTD by VU 0360172 was prevented when the animals were co-treated with MTEP (3 mg/kg, i.p., 1 h pre-LFS, hash) (VU + MTEP: 81.4 ± 2.7%, n = 5; P < 0.05 compared with VU group). (b) Summary of the mean EPSP amplitude data in (a). (c) Local injection of a relatively high dose (2 nmol, i.c.v., triangle) of the GluN2B antagonist Ro 25-6981 prevented LTD induction by LFS-300 in the presence of VU 0360172 (asterisk), whereas systemic treatment (12 mg/kg, i.p., hash) did not (i.c.v. Ro 25-6981: 97.8 ± 3.0%, n = 5, P > 0.05 compared with baseline; i.p. Ro 25-6981: 51.4 ± 3.3%, n = 5, P < 0.05 compared with baseline and between groups; two-way ANOVA RM-Sidak). (d) Summary of the mean EPSP amplitude data in (c). *P < 0.05. Calibration bars: vertical, 1 mV; horizontal, 10 ms.

Figure 6

Necessity for NMDAR ion channel function in the induction of LTD in vivo. (a) A relatively high dose of locally-injected use-dependent NMDAR ion channel blocker MK-801 (60 nmol, i.c.v., triangle), in comparison with the control (−) stereoisomer, strongly inhibited LTD induction in vivo (i.c.v. MK-801: 89.5 ± 1.3%, n = 5; i.c.v.(−)-MK-801: 52.8 ± 3.7%, n = 5, P < 0.05 compared with baselines and between groups; two-way ANOVA RM-Sidak). (b) Summary of the mean EPSP amplitude data in (c). *P < 0.05. Calibration bars: vertical, 1 mV; horizontal, 10 ms.