Developmental remodelling of the lemniscal synapse in the ventral basal thalamus of the mouse

Abstract

Synapse elimination occurs throughout the nervous system during development, and is essential for the formation of neural circuits. The mechanisms underlying synapse elimination in the brain, however, remain largely unknown. Using whole-cell patch-clamp recording in a slice preparation, we examined synaptic refinement at the somatosensory relay synapse (lemniscal synapse) in the ventral basal thalamus of the mouse during postnatal development. At 1 week old, each neuron in the ventral basal thalamus is innervated by multiple lemniscal fibres, as revealed by multiple increments of the synaptic response. By 16 days after birth (P16), the majority of neurons showed an all-or-none response, suggesting a single fibre innervation. In addition to synapse elimination, extensive modifications in synaptic properties occur during the second week after birth. The ratio of AMPA to NMDA component of the synaptic current tripled between P7 and P17. The decay constant of the NMDA component decreased by about 70% between P7 and P17; ifenprodil (3 microm) reduced the NMDA component by about 40% in neurons at P7-9, but was much less effective at P20-24. On the other hand, there was little change in the inward rectification of AMPA component between P11 and P24. Paired-pulse ratios, measured at -70 and +40 mV, were stable between P7 and P24. Whisker deprivation from P5 through P19 had no effect on the elimination or the maturation of the lemniscal synapse. These results suggest that the lemniscal synapse in the ventral basal thalamus undergoes extensive refinement during the second week, and that sensory experience has a rather limited role in this process.

Figure 1. A slice preparation of the ventral basal thalamus of the mouse

A, a sagittal section of a P24 mouse brain stained with haematoxylin nuclear counterstain. Whole-cell recording (Rec) was made in the ventral posteromedial nucleus (VPm), and stimuli (Stim) were delivered to the medial lemniscus (ML) via a concentric bipolar electrode. Abbreviations: ic, internal capsule; LV, lateral ventricle; Po, posterior thalamic nucleus; Rt, reticular thalamic nucleus; VL, ventrolateral thalamic nucleus; VPL, ventral posterolateral thalamic nucleus; ZI, zone incerta. B, a VPm neuron from a P14 mouse labelled with biocytin. C, EPSCs recorded from a VPm neuron from a P17 mouse. d-APV (100 μm, traces in red) had little effect on the peak amplitude at −70 mV, but abolished the slow component at +40 mV. The fast component at −70 or +40 mV was blocked by NBQX (10 μm). The stimulus intensity was at 100 μA. Each trace was the average of 4–5 consecutive responses. In this and other figures, stimulation artifacts were reduced first through subtractions using traces obtained with subthreshold stimuli; the remaining artifacts were truncated or removed digitally.

Figure 2. Elimination of functional inputs during development

A–D, left panels show membrane current in response to stimuli with a range of intensity in VPm neurons at P7, P12, P17 and P21, respectively; in the right panels, the peak current was plotted versus stimulus intensity. At P7 and P12, EPSCs were incremental, with more steps seen at P7 than at P12. At P17 and P21, cells mostly showed all-or-none responses. E–H, the distributions of cells with different number of lemniscal inputs over the four age groups. The distribution of the P7–9 group is significantly different from that of P16–17 or that of P20–24 (P < 0.001, K-S test), but not from that of P11–13 (P > 0.03). The distributions of the groups P16–17 and P20–24 are not significantly different from each other (P > 0.5).

Figure 3. Developmental change in AMPA/NMDA ratio

A, EPSCs recorded at +40 and −70 mV from a neuron aged P7. B, EPSCs recorded at +40 and −70 mV from a neuron aged P21. C, histogram of AMPA/NMDA ratio over the four age groups. The mean ratio was 0.79 ± 0.08 (n = 22) at P7–9, 1.79 ± 0.09 (n = 18) at P11–13, 2.44 ± 0.27 (n = 12) at P16–17, and 2.41 ± 0.31 at P20–24. The P7–9 group is significantly different from the others (P < 0.05, one-way ANOVA).

Figure 4. Developmental changes in the amplitudes of AMPA- and NMDA-EPSCs

A, histogram of the maximal EPSC recorded at −70 mV (AMPA-EPSCs) over the four age groups. The mean amplitude was 408 ± 89 pA (n = 19) at P7–9, 1002 ± 137 pA (n = 16) at P11–13, 1113 ± 139 pA (n = 11) at P16–17, and 944 ± 105 pA (n = 17) at P20–24. The P7–9 group is significantly different from the others (P < 0.05, one-way ANOVA). B, histogram of the maximal EPSC recorded at +40 mV (NMDA-EPSCs) over the four age groups. The P7–9 group is significantly different from the P16–17 and P20–24 groups (P < 0.05, one-way ANOVA). C, histogram of the first EPSC recorded at −70 mV (the minimal AMPA-EPSC) over the four age groups. The P7–9 group is significantly different from the others (P < 0.05, one-way ANOVA). D, histogram of the first EPSC recorded at +40 mV (the minimal NMDA-EPSC) over the four age groups. The four groups are not significantly different from each other (P > 0.1, one-way ANOVA).

Figure 5. Developmental decline in the decay rate of NMDAR-mediated EPSCs

A and B, EPSCs recorded at +40 mV from neurons at P8 and P20, respectively. C is the normalized result of A and B. EPSC at P20 decays much faster than that at P8. D, histogram of EPSC decay constant at +40 mV over the four age groups. The P7–9 group is significantly different from the others (P < 0.01, one-way ANOVA).

Figure 6. Developmental decline in the sensitivity to ifenprodil

A and B, effects of ifenprodil (3 μm) on EPSCs recorded at +40 mV in neurons at P7 (A) and P20 (B). NBQX (5 μm) was present throughout the recording. C, histogram of the effect by ifenprodil over the two age groups. The inhibition by ifenprodil was stronger at P7–9 than at P20–23 (P < 0.01, unpaired t test).

Figure 7. Inward rectification of AMPAR-mediated EPSCs during development

A, left panel shows traces of EPSCs recorded at various holding potentials from a neuron at P12. d-APV (50 μm) was present throughout recording to block NMDAR. The right panel shows the I–V relationship of EPSCs with a line fitted for the points from −70 to 0 mV. The points at positive holding potentials fall below the line, indicating an inward rectification. The junction potential (+5 mV) was corrected. B, results obtained from a neuron at P24. C, comparison between two age groups, P11–13 and P20–24. The rectification index is estimated as the ratio of EPSC amplitude at +35 mV to that at −35 mV in the same cell (1 means no rectification, and 0 means total inward rectification). There is no difference between the two groups (P > 0.5, unpaired t test).

Figure 8. Paired-pulse ratios of EPSCs were stable over the period from P7 to P24

A and B, paired-pulse responses of EPSCs recorded at +40 and −70 mV from neurons at P7 and P21, respectively. C, histogram of paired-pulse ratios at +40 mV (in grey) and −70 mV (hatched) over the four age groups. The means are not significantly different (P > 0.1, one-way ANOVA).

Figure 9. Lack of effect of whisker deprivation on synaptic remodelling

A, distribution of cells with the number of functional inputs for normal and whisker-deprived groups. Both groups were examined at P20–24. The result of normal mice was obtained from the same data set as in Fig. 2H. The two distributions are not significantly different from each other (P > 0.5, K-S test). B, histogram of paired-pulse ratios obtained at −70 and +40 mV for normal (hatched) and whisker-deprived group (grey). The results obtained from normal mice are not significantly different from those of whisker-deprived (P > 0.01, unpaired t test). C, histogram of EPSC amplitude obtained at −70 and +40 mV for the control (hatched) and whisker-deprived group (grey). The means are not significantly different between the two groups (P > 0.1, unpaired t test).