Functional properties of extrasynaptic AMPA and NMDA receptors during postnatal hippocampal neurogenesis

Abstract

In the mammalian hippocampus, new granule cells are continuously generated throughout life. Although it is well known that they rapidly form several thousand new glutamatergic synapses, the underlying mechanisms are not well understood. As extrasynaptic NMDA receptors are believed to support the generation of new spines, we have studied the functional properties of extrasynaptic ionotropic glutamate receptors in newborn granule cells in juvenile rats during and after synaptic integration. Using the fast application of glutamate to outside-out membrane patches, we show that all immature granule cells express functional AMPA and NMDA receptors. The density of AMPA receptors was small in cells starting to receive excitatory synaptic input (∼30 pS μm(-2)) but substantially increased during synaptic integration to finally reach ∼120 pS μm(-2) in fully mature cells. Interestingly, AMPA receptors showed a biphasic change in desensitization time constant which was slowest during synaptic integration and substantially faster before and afterwards. This was paralleled by a change in the non-desensitizing current component which was maximal during synaptic integration and about 50% smaller afterwards. Surprisingly, the NMDA receptor kinetics and density in young cells was already comparable to mature cells (∼10 pS μm(-2)), leading to an enhanced NMDA/AMPA receptor density ratio. Similar to somatic outside-out patches, iontophoretic application of glutamate onto dendrites also revealed an enhanced dendritic NMDA/AMPA ratio in young cells. These data indicate that prolonged AMPA receptor currents in newly generated young granule cells might support the effective activation of extrasynaptic NMDA receptors and therefore constitute a competitive advantage over mature cells for new synapse formation.

Figure 1. Electrophysiological classification of developing hippocampal granule cells

A, voltage responses of representative young, intermediate and mature granule cells to 1 s current injections (current steps of 5 pA for young cell, 20 pA for intermediate cell, 50 pA for mature cell). On the right, single action potentials are shown with enlarged scaling. B, scatter plot of rheobase against input resistance with mature cells (<400 MΩ) in green, intermediate cells (400 MΩ < R_in < 1 GΩ) in yellow, and young cells (>1 GΩ) in red. C, scatter plot showing the maximal rate of action potential rise plotted against input resistance. D, cumulative histogram of maximal rate of rise for young, intermediate and mature cells.

Figure 2. Increase in mEPSC frequency and amplitude during synaptic integration

A, example traces of 4 s miniature EPSC recordings for young, intermediate and mature cells, respectively. Representative single event from same cells magnified on the right. Frequency and amplitude of mEPSCs are smaller in young than in mature cells. B, scatter plots of frequency and mean amplitude of spontaneous mEPSCs against input resistance. C, scatter plots of frequency and amplitude of elicited mEPSCs during puff application of hypertonic solution (500 mm sucrose). All experiments were carried out with 0.5 μm TTX, 25 μm d-AP5 and 1 μm gabazine in bath solution.

Figure 3. AP firing matures in parallel with synapse formation

A, bar graph showing mEPSC frequency for very young cells (n= 24; R_in > 1 GΩ), young cells firing rapid APs (n= 16; R_in > 1 GΩ), intermediate cells (n= 15; 400 MΩ < R_in < 1 GΩ) and mature cells (n= 15; R_in < 400 MΩ). B, scatter plot of mEPSC frequency against maximal rate of AP rise. Linear regression shows that the development of AP properties correlates with synaptic integration (R²= 0.41, P < 0.0001). C, bar graph showing mean mEPSC amplitude for same cells as in A. D, scatter plot of mean mEPSC amplitude against maximal rate of AP rise. Filled diamonds represent mean values of young (red), intermediate (yellow) and mature (green) cells. **P < 0.01, ***P < 0.001.

Figure 4. Increase of AMPAR-mediated currents during development

A, current traces of a young, intermediate and mature cell at a holding potential of −80 mV in response to 100 ms pulses of 1 mm glutamate in the presence of 50 μmd-AP5. B, scatter plot of peak amplitude of AMPAR-mediated currents against input resistance (n= 114). C, histogram of AMPAR current density for very young cells (n= 13), young cells firing rapid APs (n= 9), intermediate (n= 9) and mature cells (n= 7). AMPAR density increases with functional maturation. D, scatter plot of peak amplitude of AMPAR-mediated current against maximal rate of rise. **P < 0.01.

Figure 5. Slow AMPAR gating kinetics during functional maturation

A, APs of a young (red, 5.91 GΩ) and a mature granule cell (green, 0.28 GΩ; left). Superimposed current traces of the same cells in response to 100 ms pulse of 1 mm glutamate (right). Current traces were fitted with a biexponential decay plus constant offset (black dashed lines) and normalized to peak amplitude. Left, insets showing young and mature morphology are taken from Schmidt-Hieber et al. . B, amplitude-weighted decay time constant of AMPAR-mediated currents for very young cells (n= 34), young cells firing rapid APs (n= 20), as well as intermediate (n= 24) and mature cells (n= 20). C, scatter plot of amplitude-weighted decay time constant against maximal rate of AP rise shows time course of desensitization during functional maturation. Dashed green line indicates mean value of mature cells. D, histogram of non-desensitizing component of AMPAR-mediated current for same set of cells as in B. E, scatter plot of the non-desensitizing component against maximal rate of AP rise. Steady state component shows a similar biphasic development as the decay time constant. *P < 0.05, ***P < 0.001.

Figure 6. Large NMDAR-mediated currents in young granule cells

A, current responses of a very young (<50 V s^-1) and a mature cell to 5 ms glutamate pulses at a holding potential of +40 mV. Extracellular solution containing 1 mm glutamate, 10 μm glycine, 10 μm CNQX. B, same cells as in A, showing NMDAR- and corresponding AMPAR-mediated currents to illustrate the different NMDAR to AMPAR current amplitude ratio. C, absolute peak amplitude of the NMDAR-mediated currents in very young cells (n= 7), young cells firing fast APs (n= 10), intermediate (n= 10) and mature cells (n= 11). D, peak current density was normalized to patch-surface area for very young (n= 6), young cells firing rapid APs (n= 3), intermediate (n= 5) and mature cells (n= 5). E, decay time constant fitted to NMDAR-mediated currents from the same cells as in C. F, the NMDAR/AMPAR current amplitude ratio is largest in very young (<50 V s^-1, n= 10) and decreases gradually with increasing maturation in young (n= 7) and intermediate cells (n= 10) towards mature values (n= 11). ***P < 0.001.

Figure 7. Enhanced NMDA/AMPA ratio in developing granule cell dendrites

A, confocal image of a newly generated young granule cell which was filled with 50 μm Alexa Fluor 594 during whole-cell recording. The application pipette (red) was filled with 150 mm glutamate and 50 μm Alexa Fluor 594 to allow positioning of the pipette tip on a dendritic shaft about 70–100 μm from soma. B, application of brief pulses of glutamate (0.2 ms) onto dendrites of a young (left) and a mature granule cell (right) evoked typical AMPAR- and NMDAR-mediated currents at –80 and +40 mV, respectively, similar to outside-out patches. C, summary plot, showing the decay time constant of NMDAR currents in dendrites of young granule cells (n= 9), which does not differ from the decay in intermediate (n= 5) and in mature cells (n= 9). D, NMDAR to AMPAR conductance ratio is significantly larger in dendrites of young granule cells (n= 9) than in mature cells (n= 9, P < 0.001). Intermediate cells already have a reduced NMDA/AMPA conductance ratio (n= 5). ***P < 0.001.

Figure 8. Enhanced NMDAR/AMPAR density during functional synaptic integration

A, scatter plot of AMPAR conductance density versus maximal rate of rise, showing a strong increase with functional maturation. B, NMDAR/AMPAR conductance ratio is high in immature cells and decreases during functional maturation towards a mean steady state value of ∼0.11. C, scatter plot of AMPAR amplitude against spontaneous EPSP (sEPSP) frequency shows increase of AMPAR-mediated current amplitude with increasing synaptic integration. D, plot of non-desensitizing component of somatic AMPAR current against sEPSP frequency. The data suggest that the non-desensitizing component during the initial phases of synaptic integration is about ∼2 times larger than in mature values. E, scatter plot of NMDAR conductance density against frequency of spontaneous EPSPs. Linear regression indicates that the density is already mature-like during the initial phase of synaptic integration. F, NMDAR/AMPAR conductance ratio is initially high and decreases with increasing synaptic integration. Dashed green lines indicate mean values for mature cells in all panels. Diamonds in C and F represent mean values of young (red), intermediate (yellow) and mature (green) cells.