Glutamatergic inputs and glutamate-releasing immature inhibitory inputs activate a shared postsynaptic receptor population in lateral superior olive

Abstract

Principal cells of the lateral superior olive (LSO) compute interaural intensity differences by comparing converging excitatory and inhibitory inputs. The excitatory input carries information from the ipsilateral ear, and the inhibitory input carries information from the contralateral ear. Throughout life, the excitatory input pathway releases glutamate. In adulthood, the inhibitory input pathway releases glycine. During a period of major developmental refinement in the LSO, however, synaptic terminals of the immature inhibitory input pathway release not only glycine, but also GABA and glutamate. To determine whether glutamate released by terminals in either pathway could spill over to activate postsynaptic N-methyl-d-aspartate (NMDA) receptors under the other pathway, we made whole-cell recordings from LSO principal cells in acute slices of neonatal rat brainstem bathed in the use-dependent NMDA receptor antagonist MK-801 and stimulated in the two opposing pathways. We found that during the first postnatal week glutamate spillover occurs bidirectionally from both immature excitatory terminals and immature inhibitory terminals. We further found that a population of postsynaptic NMDA receptors is shared: glutamate released from either pathway can diffuse to and activate these receptors. We suggest that these shared receptors contain the GluN2B subunit and are located extrasynaptically.

Figure 1. Glutamate spillover between AVCN and MNTB inputs to LSO

A: Sequential protocol S. After an LSO principal cell with a glutamatergic response in both pathways was located, the NMDAR-mediated component was pharmacologically isolated, and then the experiment began. Aa: Representative raw (gray) and average (black) current traces (P3 cell), showing the mixed (‘no drug’) GABA/glycine/glutamate traces in MNTB and the glutamate traces in AVCN before the beginning of the experiment, the Baseline NMDAR responses after application of picrotoxin (Ptx), strychnine (Str), and CNQX, and the decrementing Test responses in the presence of MK-801. Ab: Schematic showing the spillover experiment on pharmacologically isolated NMDAR-mediated response, for the traces shown above. Baseline NMDAR responses (“Bsl”) were obtained in both pathways (interleaved), MK-801 perfusion began (gray area), and Test responses to P₁ (MNTB-LSO in this instance) and then to P₂ (AVCN-LSO in this instance) were collected. B: The amplitude of the first Test EPSC (T₁) in MK-801 was reduced more in P₂ than in P₁ for all 10 cells recorded in this manner (T₁ EPSC amplitude as % of Baseline: P₁ 84.9±3.6%, P₂ 35.7±3.7%; P=0.002, W=55, N=10, Wilcoxon). Each pair of adjoined symbols represents one cell, red indicates that MNTB was the first pathway stimulated (P₁ = MNTB-LSO) and green that AVCN was the first pathway stimulated (P₁ = AVCN-LSO). C: Interleaved protocol I. Ca: Representative raw (gray) and average (black) current traces (P3 cell) for Baseline EPSCs and for Test EPSCs after the further addition of MK-801. Cb: Schematic of protocol I, for the traces shown above. Both Baseline (“Bsl”) and Test responses were collected in an interleaved manner, and the period of MK-801 exposure before T₁ was nearly identical for the two pathways. D: T₁ amplitude in MK-801 was reduced more in P₂ than in P₁, for all cells recorded using protocol I (T₁ EPSC amplitude, % Baseline: P₁ 79.3±3.9%, P₂ 52.6±7.9 %; P=0.0298, W=32, N=8, Wilcoxon). E: Responses to first test in P₁ (P₁T₁) were similar for 10- (S) and 15-minute (I) MK-801 incubations (P = 0.303, U=28, Mann-Whitney). Note: data re-plotted from Figures 1B and D. F: Responses to first test in P₂ (P₂ T₁) were reduced by a larger amount in the sequential protocol (S) than in the interleaved protocol (I) (P=0.0414, U=20, Mann Whitney). Note: data re-plotted from Figures 1B and D. G: Response amplitude in P₁ decremented more quickly when stimuli were interspersed with stimuli to opposite pathway (I), than when P₁ alone was stimulated (time-constant τ: S 6.7 ± 1.3 trials, I 3.3 ± 0.5 trials; P=0.0273, U=18, Mann Whitney). All experiments in 0 Mg⁺⁺ ACSF, at room temperature.

Figure 2

Glutamate spillover occurs at physiologically relevant temperatures. A: Representative raw (gray) and average (black) traces for Baseline and Test responses at physiologically relevant temperatures (34–36C), in sequential protocol (P7 slice), 0 Mg⁺⁺ ACSF, −70mV holding potential. B: Test EPSCs were reduced significantly in P₂ relative to P₁ (P1T1 85.6 ± 3.4%; P2T1 33.5 ± 3.3%; P=0.0010, W=66, N=11, Wilcoxon). Each pair of points corresponds to a single neuron, with red and green indicating P₁ as above. C: Under Protocol S, T₁ amplitudes in P₂ were equally decremented at room temperature and at physiological temperatures (P=0.7513, U=50, Mann Whitney). D: T₁ amplitudes in P₁ did not differ significantly between pathways (MNTB: 97.2 ± 6.9%, N=6; AVCN 97.7 ± 4.0, N=6; P=0.6991, U=15, Mann Whitney). Ea: Raw (gray) and average (black) traces for interleaved protocol shown in Eb, performed with no MK-801 in the perfusate (P8 slice). Gray symbols indicate the values used for measuring Baseline and Test amplitudes (last 20 test stimuli). F: In the absence of MK-801, Test responses (last 20 stimuli) were not substantially decremented relative to baseline in either the AVCN or the MNTB pathway (MNTB 94.1 ± 8.0; AVCN 83.6 ± 6.0; N=9, P=0.4894, U=32, Mann Whitney). All experiments in 0 Mg⁺⁺ ACSF, at physiological temperature.

Figure 3

Glutamate spillover occurs in normal Mg⁺⁺ concentrations, and at physiologically relevant stimulation intensities. A: Example cell (P5), Baseline and Test responses in Mg⁺⁺-containing ACSF, at holding potential of +40mV. B: For 4 cells, in normal Mg⁺⁺-ACSF, using sequential protocol, T₁ EPSC is reduced more in P₂ than in P₁ (P1T1 92.0 ± 1.5; P2T1 42.9 ± 9.7; P=0.0286, U=1, N=4, Mann Whitney). C: Recording from synaptically connected MNTB-LSO cell pair, before addition of strychnine, picrotoxin and CNQX; P3 slice. Current injection at MNTB cell (top) elicits an action potential that produces a postsynaptic potential in LSO principal cell (bottom). D: Representative minimal stimulation experiment, sequential protocol performed after establishing minimal stimulation in both pathways. P8 slice, recordings in 0 Mg⁺⁺ ACSF, at −70mV. Da: 50 Baseline responses were collected in each pathway, MK-801 was applied for 10 minutes, and then stimulation began again in P₁. Db: Raw (gray) and average (black) responses for the cell in a. E: EPSC amplitude histograms for Baseline and Test periods in Path 1 (top) and Path 2 (bottom), for the cell shown in C. F: Population data from minimal stimulation experiments to ensure that single fibers were stimulated in both pathways. Solid horizontal line indicates mean failure ‘amplitude’ and dashed horizontal lines indicate mean ± 2 standard deviations, for each pathway. (EPSC amplitude, as % Baseline: P₁T₁ 98.7 ± 9.6; P₂T₁ 60.0 ± 3.3; P=0.0078, W=36, N=8, Wilcoxon). All experiments at physiological temperature; AB in Mg⁺⁺-containing ACSF; C–F in 0 Mg⁺⁺ ACSF.

Figure 4

Decrement in P₂ is not due to inadvertent prior stimulation of P₂. A: Example cell (P6), response to 10-pulse train delivered to MNTB-LSO, in the presence of picrotoxin, strychnine, and cyclothiazide (Ctz), holding potential −70mV (upper left). Paired-pulse ratio (PPR) for MNTB pathway alone (Response 2/Response 1) for this cell: 0.50 ± 0.03. Upper right: Response to 10-pulse train delivered to AVCN-LSO (PPR for AVCN pathway alone 0.48 ± 0.02). Lower left: Response to 10-pulse MNTB-LSO train preceded by a 10-pulse train to the AVCN-LSO pathway. PPR for MNTB stimulation preceded by AVCN stimulation (Response 12/Response 11) for this cell is 0.50 ± 0.05). Lower right: Response to 10-pulse train AVCN-LSO train preceded by a 10-pulse train to the MNTB-LSO pathway (PPR = 0.51 ± 0.02). Ba: PPRs for MNTB alone do not differ from PPRs measured immediately after AVCN trains, (MNTB alone 0.76 ± 0.09; AVCN≫MNTB 0.77 ± 0.06; P=1.0, U=32, N=8, Mann Whitney). Bb: PPRs for AVCN alone do not differ from PPRs preconditioned by MNTB trains (AVCN alone 0.74 ± 0.06; MNTB≫AVCN 0.76 ± 0.06, P=0.6991, U=15, Mann Whitney). C: Baseline and Test EPSCs for cell shown in A. D: Test EPSCs (T₁) were reduced more in P₂ than in P₁ (P₁T₁ 96.5 ± 4.9%, P₂T₁ 49.1 ± 6.2%; P=0.0143, U=0.0, Mann Whitney). All experiments in Mg⁺⁺-containing ACSF, at physiological temperature.

Figure 5

AVCN-LSO and MNTB-LSO pathways share NMDA, but not AMPA, receptors. A: Left: AMPAR-mediated response recorded at −70mV for representative cell (P5), showing individual responses to AVCN-LSO and MNTB stimulation, and response to joint stimulation of both pathways. Right: For the same cell, with CNQX in the bath, joint stimulation yields an NMDAR-mediated response only slightly larger than the sum of responses in each pathway. B: Left: For 11 cells, joint response amplitude (to stimulation of both pathways) equals summed amplitudes from stimulation of AVCN-LSO and MNTB individually (MNTB 32.3 ± 2.0%, AVCN 69.2 ± 2.9%; joint 102.2 ± 3.8%). Right: for the same neurons, the NMDAR-mediated response to joint stimulation is smaller than the sum of the individual responses (MNTB 32.1 ± 3.2%, AVCN 67.7 ± 3.5%; joint 73.6 ± 3.9%). All experiments in Mg⁺⁺-containing ACSF, at physiological temperature.

Figure 6

NMDAR-mediated EPSC exhibits a slower, more variable rise time in MNTB than in AVCN pathway. A: Rise times for isolated NMDAR-mediated response in MNTB-LSO and AVCN-LSO pathways, measured as the time (in ms) between 10% and 90% of peak response. (rise-time for MNTB: 8.3 ± 7.0 ms; AVCN 6.3 ± 4.2ms, P=0.0338, U=117, Mann Whitney, N=20) B: Coefficients of variation of rise times, for cells shown in A (MNTB 0.35 ± 0.03; AVCN 0.22 ± 0.01; P = 0.0106, U=105, Mann Whitney) C: Representative traces from a P5 cell, held at +40mV. After application of strychnine, picrotoxin and CNQX to isolate the NMDAR-mediated EPSC, the further application of Ro-25 reduces response amplitude equally in both pathways (for this cell, EPSC amplitude as % Ctrl: MNTB 44.8 ± 4.5; AVCN 37.5 ± 3.4). D: In MNTB-LSO pathway, rise times are shortened by application of Ro-25 (Bsl 9.1 ± 1.1ms, Ro-25 5.0 ± 0.5ms; P=0.0078, W=43, N=9, Wilcoxon). E: Rise times in the AVCN-LSO pathway are unaffected by Ro-25 (Bsl 7.1 ± 0.4ms, Ro-25 6.9 ± 0.4ms; P=0.4961, W= 13, N=9, Wilcoxon). All experiments at physiological temperature.