NMDA EPSC kinetics do not regulate the critical period for LTP at thalamocortical synapses

Alison L. Barth and Robert C. Malenka

Nancy Friend Pritzker Laboratory, Department of Psychiatry and Behavioral Sciences, 1201 Welch Road, Room P105, Stanford University School of Medicine, Palo Alto, California 94304-5485, USA

Correspondence should be addressed to A.B. ()

Methods

Animals. C57bl/6J mice aged P3–27 were used for this study, where P0 indicates the day of birth. Slices were cut according to the methods described previously1 with the following modifications to accommodate the changing anatomy of the neonatal mouse brain. From P3–P6, slices were cut at an angle 50° to the midline on a flat surface. After P7, we found that slices retained better connectivity if we cut them at 50° to the midline on a 10° incline, with the olfactory bulbs facing downward. At P3–5, our 400-M-thick slices gave approximately one slice per brain where we could elicit a significant (>100 V) field potential in layer IV using thalamic stimulation. After P8, we could sometimes find two slices per animals that gave us a robust field response. For animals aged P3–13 slices were cut in Ringer’s solution containing 119 mM NaCl, 2.5 mM KCl, 1.3 mM MgSO4, 1 mM NaH2PO4-H20, 26.2 mM NaHCO3, 2.5 mM CaCl2-2H2O, 11 mM glucose. In older animals (> P17), slices were cut in Ringer’s solution with 1 mM kynurenic acid (Sigma, St. Louis, Missouri) because we found this improved the viability of the slices for whole-cell recording.

Whole-cell recording. Slices were continuously perfused with Ringer’s equilibrated with a 95/5% mixture of O2 and CO2. Before whole cell recording was attempted, we optimized the field potential response stimulating in the thalamus and recording in layer IV of barrel cortex. Typically, stimulating from one location in the thalamus would yield a 100 V or greater amplitude field potential response in an area in S1 covering approximately 2–5 barrels. Glass recording electrodes (Garner Glass Company, internal dimension 1.0 mm) were pulled to 3–6 M resistances. Internal recording solution was composed of 117 mM Cs-gluconate, 20 mM HEPES, 0.4 mM EGTA, 2.8 mM NaCl, 5 mM tetraethylammonium chloride with 4 mM MgATP and 0.3 mM NaGTP added fresh to each day’s aliquot. Internal solution was brought to a pH of 7.25–7.35 and 270–290 mOsm. Recordings were made from cells in layer IV, as determined by a dark band of dense, granular-appearing cells in very young animals (P3–4) which slowly evolved into more distinct barrel-like structures after P5. Input and series resistance were monitored continuously online using a small (3–5 mV) hyperpolarizing voltage step.

To verify that we were indeed recording monosynaptic thalamocortical EPSCs, we routinely applied a 1 Hz stimulation train to look for changes in latency of the response, reasoning that polysynaptic responses were likely to exhibit variable EPSC latency. Cells that showed changes in latency or a substantial increase in failures (more than 20%) were excluded from further analysis. Furthermore, for all experiments where we examined NMDAR EPSC kinetics, we used the AMPA/kainate antagonist NBQX that should eliminate any EPSC resulting from a polysynaptic input since synaptic activation of NMDARs alone is unlikely to be sufficient to elicit action potentials and drive polysynaptic responses. For several reasons, it is also very unlikely that we were antidromically activating corticothalamic pathways that might generate a monosynaptic response from axon collaterals to layer IV. First, previous experiments in rat using voltage-sensitive dyes showed that thalamic stimulation in this preparation resulted in a strong depolarization in layer IV but not in deep cortical layers. Second, thalamic DiI labeling in our slice preparation revealed labeling of barrel-like structures in layer IV but not of cell bodies in deep layers in the barrel cortex, as would be expected if this slice contained a significant intact corticothalamic projection. This indicates that the angle at which we cut these slices preserves thalamocortical connections but disrupts descending corticothalamic projections.

With regard to the possible contribution of inhibitory postsynaptic currents (IPSCs) to the synaptic responses we examined, this is unlikely to be to be an issue until roughly P10 and later ages. The maturation of the inhibitory system has been well described by others2, and their results confirm our observations that IPSCs were rarely observed until after P10. On the occasions that we did isolate an IPSC, this was immediately apparent by a large outward current during pairing at 0 mV, since IPSCs reverse at holding potentials greater than –70 mV. Such cells were not included in our analysis. For measurement of NMDAR EPSCs, all recordings were done in NBQX to eliminate any AMPA/kainate components of the response and this had the added benefit of eliminating any IPSCs which must, by definition, be polysynaptic since ascending thalamic inputs are exclusively glutamatergic.

Drugs were applied in the superfusing solution in the following concentrations: ifenprodil (Tocris) 3 M; 2,3-Dioxo-6-nitro-1,2,3,4-tertrahydrobenzo[f]quinoxaline-7-sulphonamide (NBQX; Tocris) 10 M; picrotoxin (Sigma) 100 M; and D-APV (Tocris) 50 M.

Data Analysis. Time constants of decay were fitted to NMDAR EPSCs (average of 30 traces) with Origin software using a single exponential decay function. We believe this is adequate to describe the kinetics of NMDAR EPSCs for several reasons. First, in their original report describing activity-dependent changes in NMDAR kinetics, Carmignoto and Vicini3 state that NMDAR EPSCs were well-fit by a single exponential in visual cortex of animals aged 9–14 days. We also found this to be the case. Indeed, when we used both a single and a double exponential to fit our NMDAR EPSCs, we found that the R2 value, a measurement of curve fitting, was only marginally better with the double exponential. We only used time constants of decay if the R2 value was better than 0.90; typically (more than 90 percent of cases) this value was 0.95 or better. Although we do include some values from animals at P18 and P27, these EPSCs were also well fit by a single exponential (again, data points were only included if the R2 value was >0.90). Thus, although it might be more appropriate to develop a weighted time constant for NMDAR EPSCs from older animals, for comparison’s sake we used the same method as for the younger animals. In any case, the absolute value of the time constant is not critical to our conclusions; the important point is that by any criterion, the decay of the NMDAR EPSC is very slow in young animals and becomes fast long after the critical period for LTP ends.

To provide a simple visual comparison of NMDAR EPSCs over time, we scaled traces from individual cells to 100 pA and then averaged them according to postnatal day. These averaged traces from a large number of individual cells were then used to estimate changes in the charge transfer over the first postnatal weeks by simply integrating the area under the traces (Origin software).

Discussion

Our data raise a number of interesting points that space constraints precluded us from fully discussing in our Brief Communication. We will address two issues of potential interest to others who work in this area: first, the possibility that there are large changes in the magnitude of the NMDAR EPSC during early development that might result in a net decrease of Ca2+ entry into the cell during receptor activation, and second, the population of ifenprodil-insensitive NMDARs and the properties of their decay kinetics.

Magnitude of the NMDAR EPSC during early development

In this paper, we present evidence that changing decay kinetics of the NMDAR EPSC do not provide a convincing explanation for the duration of the critical period for LTP induction at thalamocortical synapses in the mouse barrel cortex. Although we find that NMDAR EPSCs do become faster over early postnatal development, this transition occurs days to weeks after the critical period for LTP induction ends. The duration of NMDAR channel opening is associated with the amount of Ca2+ that can enter the cell, and it has been widely hypothesized that the triggering of LTP requires a large rise in intracellular Ca2+ concentration beyond some critical threshold 4. Therefore, it is formally possible that although the decay kinetics of the NMDAR EPSCs at thalamocortical synapses do not change substantially during the window surrounding the critical period for the triggering of LTP, the absolute magnitude of the NMDAR EPSC might decrease, such that the amount of Ca2+ that can enter the cell is below the threshold for LTP induction. It is important to remember, however, that according to current thinking4,5 the key variable in triggering LTP is the magnitude of the rise in Ca2+ at individual synapses (i.e. individual dendritic spines). Simply measuring the size of the evoked NMDAR EPSC, which reflects activity at some unknown number of synapses, does not provide information about the number and properties of NMDARs at individual synapses. To obtain such information likely requires recording miniature NMDAR EPSCs (i.e. the NMDAR EPSCs generated by individual synapses), an experimentally challenging endeavor which we have not been able to perform successfully in layer IV cells of mouse barrel cortex.

Decay kinetics of ifenprodil-resistance NMDAR EPSCs

When we initiated these experiments, we expected that application of the NR2B antagonist ifenprodil, which has a 400-fold greater affinity for NR1-NR2B than NR1-NR2A heteromers6, might reveal an ifenprodil-resistance population of NMDARs with fast decay kinetics. Such receptors were likely to be NR1-NR2A or NR1-NR2A-NR2B heteromers, and the presence of these receptor channels has been posited based upon a number of molecular and electrophysiological studies7–11. We were thus surprised to find that the decay kinetics of the ifenprodil-resistance NMDAR EPSCs were very long and almost indistinguishable from control NMDAR EPSCs (see accompanying figure). What feature of the NMDAR might be responsible for this effect? We speculate that the long decay kinetics of ifenprodil-resistant NMDARs may stem from several factors. First, we expect that the NMDARs mediating these residual currents may contain some NR2A subunits, and that the proportion of this subunit within the receptor channel is not sufficient to confer fast decay kinetics but does interfere with ifenprodil binding. Other investigators have described an intermediate-affinity binding site for ifenprodil that is likely to conform to such heteromers 12. Second, it is possible that a splice variant of the NR1 subunit influences NMDAR EPSC decay kinetics in a dominant manner such that in some cases the presence of NR2A is insufficient to confer fast kinetics. Finally, it is possible that an additional NR2 subunit, such as NR2D, can produce long decay kinetics even in the presence of some NR2A. Although not restricted to barrel cortex, previous in situ data7 show NR2B and NR2D expression in the neocortex during the first few postnatal weeks. To our knowledge, the properties of such NR2 heteromultimers (that is, NMDARs containing some proportion of NR1, NR2A, NR2B, and/or NR2D) have not been adequately described in the current literature.

Summary

Our results, together with those from other laboratories, provide increasing evidence to dissociate the duration of NMDAR-mediated currents from periods of enhanced synaptic plasticity. For example, in the zebrafinch, song learning can proceed even in the absence of slow NMDAR EPSCs in necessary brain nuclei13. In visual cortex of mammals, decreasing NMDAR EPSC duration has been reported to occur at the beginning of the critical period for ocular dominance plasticity10, not at the end as would be expected if it negatively regulated plasticity. Our data support a model where the NMDAR mediates developmental changes in synaptic and experience-dependent plasticity not solely through changes in its biophysical properties but also perhaps through changes in its molecular associations.

1.Agmon, A. & Connors, B. W. Thalamocortical responses of mouse somatosensory (barrel) cortex in vitro. Neuroscience41, 365–379 (1991).

2.Agmon, A. & O'Dowd, D. K. NMDA receptor-mediated currents are prominent in the thalamocortical synaptic response before maturation of inhibition. J. Neurophysiol.68, 345–349 (1992).

3.Carmignoto, G. & Vicini, S. Activity-dependent decrease in NMDA receptor responses during development of the visual cortex. Science258, 1007–1011 (1992).

4.Malenka, R. C. & Nicoll, R. A. NMDA-receptor-dependent synaptic plasticity: multiple forms and mechanisms. Trends Neurosci16, 521–527 (1993).

5.Malenka, R. C. & Nicoll, R.A. Long-term potentiation - a decade of progress? Science 285, 1870–1874 (1999).

6.Williams, K. Ifenprodil discriminates subtypes of the N-methyl-D-aspartate receptor: selectivity and mechanisms at recombinant heteromeric receptors. Mol. Pharmacol.44, 851–859 (1993).

7.Monyer, H., Burnashev, N., Laurie, D. J., Sakmann, B. & Seeburg, P. H. Developmental and regional expression in the rat brain and functional properties of four NMDA receptors. Neuron12, 529–540 (1994).

8.Sheng, M., Cummings, J., Roldan, L. A., Jan, Y. N. & Jan, L. Y. Changing subunit composition of heteromeric NMDA receptors during development of rat cortex. Nature368, 144–147 (1994).

9.Flint, A. C., Maisch, U. S., Weishaupt, J. H., Kriegstein, A. R. & Monyer, H. NR2A subunit expression shortens NMDA receptor synaptic currents in developing neocortex. J. Neurosci.17, 2469–2476 (1997).

10.Roberts, E. B. & Ramoa, A. S. Enhanced NR2A subunit expression and decreased NMDA receptor decay time at the onset of ocular dominance plasticity in the ferret. J. Neurophysiol.81, 2587–2591 (1999).

11.Nase, G., Weishaupt, J., Stern, P., Singer, W. & Monyer, H. Genetic and epigenetic regulation of NMDA receptor expression in the rat visual cortex. Eur. J. Neurosci.11, 4320–4326 (1999).

12.Kew, J. N., Richards, J. G., Mutel, V. & Kemp, J. A. Developmental changes in NMDA receptor glycine affinity and ifenprodil sensitivity reveal three distinct populations of NMDA receptors in individual rat cortical neurons. J. Neurosci.18, 1935–1943 (1998).

13.Livingston, F. S., White, S. A. & Mooney, R. Slow NMDA-EPSCs at synapses critical for song development are not required for song learning in zebra finches. Nat. Neurosci.3, 482–488 (2000).

Supplementary Figure

Ifenprodil-resistant NMDAR EPSCs remain long. Time constants of decay for NMDAR EPSCs in the presence of 3 M ifenprodil were calculated using single exponentials (Origin software) for each postnatal day. Values from each day were averaged (P4, n = 2; P5, n = 4; P6, n = 6; P7, n = 5; P8, n = 4; P9, n = 7).

1