Calcium-permeable presynaptic AMPA receptors in cerebellar molecular layer interneurones

Abstract

Axons of cerebellar molecular layer interneurones (MLIs) bear ionotropic glutamate receptors. Here, we show that these receptors elicit cytosolic [Ca2+] transients in axonal varicosities following glutamate spillover induced by stimulation of parallel fibres (PFs). A spatial profile analysis indicates that these transients occur at the same locations when induced by PF stimulation or trains of action potentials. They are not affected by the NMDAR antagonist AP-V, but are abolished by the AMPAR inhibitor GYKI-53655. Mimicking glutamate spillover by a puff of AMPA triggers axonal [Ca2+]i transients even in the presence of TTX. Addition of specific voltage-dependent Ca2+ channel (VDCC) blockers such as omega-AGAIVA and omega-conotoxin GVIA or broad range inhibitors such as Cd2+ did not significantly inhibit the signal indicating the involvement of Ca2+-permeable AMPARs. This hypothesis is further supported by the finding that the subunit specific AMPAR antagonist IEM-1460 blocks 75% of the signal. Bath application of AMPA increases the frequency and mean peak amplitude of GABAergic mIPSCs, an effect that is blocked by philanthotoxin-433 (PhTx) and reinforced by facilitating concentrations of ryanodine. By contrast, a high concentration of ryanodine or dantrolene reduced the effects of AMPA on mIPSCs. Single-cell RT-PCR experiments show that all GluR1-4 subunits are potentially expressed in MLI. Taken together, the results suggest that Ca2+-permeable AMPARs are colocalized with VDCCs in axonal varicosities and can be activated by glutamate spillover through PF stimulation. The AMPAR-mediated Ca2+ signal is amplified by Ca2+-induced Ca2+ release from intracellular stores, leading to GABA release by MLIs.

Figure 1. Stimulation of PFs evokes intracellular [Ca²⁺]_i transients in MLI axons

A, left panel, time course of ΔF/F₀ in an axonal hot spot following PF stimulation (arrow; 10 pulses, 50 Hz); simultaneous current recording (inset; stimulation duration indicated by the bar above the current trace). Note the glutamate spillover recognizable by the slow current component. Right panel, ΔF/F₀ image at the peak of the response; the analysed region (hot spot) is indicated by a white arrow. B, left panel, averaged peak ΔF/F₀ is plotted versus number of PF stimulations (N= 3 cells; multiple stimulations have been performed at a frequency of 50 Hz). Right panel, peak ΔF/F₀ is plotted versus the frequency for 5 PF stimulations (N= 3 cells). C, left panel, time course of ΔF/F₀ evoked by a 4AP train in the same hot spot as in A; simultaneous current recording (inset). Right panel, ΔF/F₀ image at the peak of the response; the analysed region (hot spot) is indicated by a white arrow. D, ΔF/F₀ image of the same axon as in A (90 deg inclination, top image). A dotted line running along the axon is superimposed. ΔF/F₀ analysis was performed along the dotted line (see Methods). ΔF/F₀ values corresponding to a 4-AP train (black trace) or PF stimulation (red trace) are plotted versus axonal distance (bottom graph). All pseudocolour calibration bars in this and in the following figures indicate ΔF/F₀ as a percentage. Space calibration bar for the images in A, C and D: 5 μm.

Figure 2. AMPARs mediate the axonal [Ca²⁺]_i transients evoked by PF stimulation

A, left panel, ΔF/F₀ time course in a MLI axon and corresponding current recording (inset) during PF stimulation (10 pulses, 100 Hz at the time indicated by the arrow). Right panel, ΔF/F₀ image at the peak of the response of the recorded axon (hot spot of interest indicated by the white arrow). B, same as in A in the presence of the specific NMDAR antagonist AP-V (50 μm). C, same as A in the presence of AP-V (50 μm) and of the specific GluR2-lacking AMPAR antagonist GYKI 53655 (40 μm). D, summary plot of pooled data. Space calibration bar for the images in A, B and C: 2 μm.

Figure 3. A puff of AMPA mimicks glutamate spillover

A, illustration of the AMPA puff protocol. GCL: granular cell layer; PCL: Purkinje cell layer; ML: molecular layer. B, upper panel, ΔF/F₀ image of a MLI soma at the peak of the response to a puff of AMPA (100 μm, 500 ms). Lower panel, time course of ΔF/F₀ for the somatic region of interest delimited by the white square on the upper panel and simultaneous current recording (inset). C, left panel, time course of ΔF/F₀ for an axonal hot spot and simultaneous current recording (inset) during an AMPA puff. Right panel, ΔF/F₀ image of the axon at the peak of the response to the puff. The hot spot analysed is indicated by the white arrow. Same cell as in B. D, normalized current peak amplitude and peak ΔF/F₀ increase as a function of AMPA concentration (▪: current; ◯: ΔF/F₀; n= 3). E, normalized current peak amplitude and peak ΔF/F₀ increase as a function of puff duration (▪: current; ◯: ΔF/F₀; n= 3). Space calibration bar for the images in B and C: 5 μm.

Figure 4. Presynaptic AMPARs are calcium permeable and colocalize with VDCC in defined hot spots

A, peak ΔF/F₀ of hot spots in MLI axons during puffs of AMPA (100 μm, 500 ms; n= 15) recorded in TTX (500 nm). Pairs of circles joined by a line represent one cell. B, summary plot of the relative peak ΔF/F₀ of the hot spots during a puff of AMPA in different conditions (500 nm TTX, n= 15; 500 nm TTX + 50 μm Cd²⁺, n= 15; 500 nm TTX + 50 μm Cd²⁺+ 40 μm CTZ, n= 7; 500 nm TTX + 40 μm IEM 1460, n= 4). C, left panel, typical current recording of a MLI during a puff of AMPA in control condition, after 2 min and 6 min of bath applied IEM 1460. Right panel, corresponding ΔF/F₀ time courses (averages of 2 or 3 trials). D, left panel, typical current recording of a MLI during a puff of AMPA in the presence of 500 nm TTX, after addition of 50 μm Cd²⁺ and CTZ (50 μm). Right panel, corresponding ΔF/F₀ time courses.

Figure 5. AMPA increases the frequency and mean peak amplitude of GABAergic mIPSCs

A, upper panels, samples of a continuous current recording of mPSCs in a MLI before (left trace) and during (right trace) the application of AMPA (0.5 μm). Lower panels, plots for the same cell of the amplitude of GABAergic mIPSCs before (left graph) and after (right graph) the application of AMPA. The points joined by a black line (panel C) correspond to the mean ±s.e.m. for this dataset. B, histograms of GABAergic mIPSC peak amplitudes in control (left) and in the presence of AMPA (right) averaged from 8 cells. The white curves corresponds to the best approximation of the data by a Gaussian function. C, mean amplitude values are shown for n= 8 cells in both conditions. D, CV for mIPSC amplitude values are shown for n= 8 cells in both conditions.

Figure 6. Ca²-permeable AMPARs in MLI-MLI terminals

Aa, increase in the frequency of mIPSCs (upward deflections) in a PC held at +30 mV following application of AMPA (0.5 μm). Ab, the same result is observed in the presence of PhTx (1 μm). Ba, increase in the frequency of mIPSCs (downward deflections) in a MLI held at −70 mV following application of AMPA (0.5 μm). Bb, this effect is blocked in the presence of PhTx (1 μm). C, effects of PhTx (1 μm) on the AMPA (0.5 μm)-induced increases of frequency of mIPSCs recorded either from Purkinje cells (n= 3) or from MLIs (control: n= 12; PhTx: n= 7) as indicated underneath the graphs. Each column represents the mean ±s.e.m.

Figure 7. Ryanodine (1 μm) amplifies AMPA effects on GABAergic mIPSCs

A, plot of the sum of the amplitude of all GABAergic synaptic events detected during 20 s sample intervals in control conditions (grey) or in the presence of ryanodine (1 μm; blank). The bar indicates the duration of AMPA (1 μm) application to the bath. B, summary of three experiments in control and in the presence of 1 μm ryanodine. C, typical recording of synaptic currents in the presence of ryanodine (1 μm) and AMPA (indicated by the bar above the current recording). Note the presence of a burst. D, effects of ryanodine (50 μm; n= 4) and dantrolene (10 μm; n= 4) on the AMPA (0.5 μm)-induced increases of mIPSC frequency recorded from MLIs (control: n= 12). E, effects of ryanodine (50 μm; rya 50; n= 4) and dantrolene (10 μm; dantro; n= 4) on the amplitude of mIPSCs in the presence of AMPA (0.5 μm; control: n= 12). For D and E, each column represents the mean ±s.e.m.

Figure 8. All AMPAR subunits are expressed in MLIs

A, result of a RT-PCR experiment showing all subunits GluR1–4 expressed in a single MLI. B, summary of 23 RT-PCR experiments: percentage of MLIs expressing GluR1–4 subunits.