Heterogeneous CaMKII-Dependent Synaptic Compensations in CA1 Pyramidal Neurons From Acute Hippocampal Slices

Abstract

Prolonged changes in neural activity trigger homeostatic synaptic plasticity (HSP) allowing neuronal networks to operate within functional ranges. Cell-wide or input-specific adaptations can be induced by pharmacological or genetic manipulations of activity, and by sensory deprivation. Reactive functional changes caused by deafferentation may partially share mechanisms with HSP. Acute hippocampal slices are a suitable model to investigate relatively rapid (hours) modifications occurring after denervation and explore the underlying mechanisms. As during slicing many afferents are cut, we conducted whole-cell recordings of miniature excitatory postsynaptic currents (mEPSCs) in CA1 pyramidal neurons to evaluate changes over the following 12 h. As Schaffer collaterals constitute a major glutamatergic input to these neurons, we also dissected CA3. We observed an average increment in mEPSCs amplitude and a decrease in decay time, suggesting synaptic AMPA receptor upregulation and subunit content modifications. Sorting mEPSC by rise time, a correlate of synapse location along dendrites, revealed amplitude raises at two separate domains. A specific frequency increase was observed in the same domains and was accompanied by a global, unspecific raise. Amplitude and frequency increments were lower at sites initially more active, consistent with local compensatory processes. Transient preincubation with a specific Ca²⁺/calmodulin-dependent kinase II (CaMKII) inhibitor either blocked or occluded amplitude and frequency upregulation in different synapse populations. Results are consistent with the concurrent development of different known CaMKII-dependent HSP processes. Our observations support that deafferentation causes rapid and diverse compensations resembling classical slow forms of adaptation to inactivity. These results may contribute to understand fast-developing homeostatic or pathological events after brain injury.

Keywords: CA1 pyramidal neurons; CaMKII; deafferentation; homeostatic synaptic plasticity; rat brain slices.

FIGURE 1

Upregulation of mEPSCs amplitude and frequency in acute slices lacking CA3. (A) Left, experimental design and representative somatic current recordings from two different neurons at early (T1) and late (T2) stages. Right, superimposed average mEPSCs waveforms (above) and rescaled traces (below). (B) Summary plots for mEPSCs amplitude, frequency, rise and decay times for all recorded neurons. Each point corresponds to the mean value from a single neuron. (C) Amplitude cumulative probability distributions of all mEPSC recorded at T1 or T2. (D) Interevent interval (IEI) distributions. Random permutation test. (E) Membrane properties and series resistance for the two cell populations (Unpaired t-test). N = 9 cells for T1 and N = 8 for T2, recorded from 6 rats (1–4 slices per rat).

FIGURE 2

Heterogeneous synaptic adaptations in different groups of synapses. All mEPSCs recorded at a time interval (T1 or T2) were pooled together and sorted by rise time (τ_r). (A) Left, distribution of mEPSC amplitude percent changes for τ_r-binned data (0.1 ms). The red curve corresponds to a bimodal fit (fast and slow-rising mEPSCs). Insets, superimposed average traces for T1 (black) and T2 (red), rise-time values indicated by the arrows. Data was separated into fast and slow rising groups using a 95% confidence interval obtained from the multi-peak fit. Right, mean amplitude percent increase for slow and fast-rising mEPSC. (B) Left, decay time (τ_d) percent change distribution for τ_r-binned data. Right, decay time change summary plot for slow- and fast-rising mEPSC. (C) Left, superimposed frequency histograms for all events recorded at T1 or T2. Right, probability density curves showing a relative increase in the incidence of fast-rising events. (D) Left, τ_r-binned frequency percent increases. Right, average frequency increase for fast and slow events. Bootstrap with Bonferroni correction [bar plots in panels (A,B,F)]. N = 9 cells for T1 and N = 8 for T2 (same data as for Figure 1). *p < 0.05.

FIGURE 3

CaMKII is involved in different adaptations in acute slices. (A) Experimental design. Slices were preincubated for 2 h with CN21 or SCR peptide and drugs were washed out for 2 h (T1 and T2 are the same as for Figures 1, 2). (B) Representative current traces recorded at T1 and T2 in each condition (C) Mean mEPSCs amplitudes at T1 and T2 (left) and overlaid cumulative probability curves for CN21 (center) and SCR (right), including all mEPSCs recorded during each period. Black curves: distributions at T1. Two-way ANOVA with Bonferroni correction for bar plot, random permutations for cumulative probability plots. (D) Same as panel (C) for mEPSCs frequency and interevent interval (IEI). (E) Percent changes in amplitude and decay time for τ_r-binned data. Bootstrap with weighted P-values. (F) Average amplitude and decay time percent changes for slow and fast-rising groups. Bootstrap with Bonferroni correction. (G) Amplitude vs. decay time raw changes for CN21 or SCR peptide. Each point corresponds to the average change for τ_r bin. Pearson correlation. (H) Frequency histograms (left) for all events recorded at T1 or T2 and probability density curve (right), in SCR condition. (I) τ_r-binned frequency percent increase. (J,K) Same as (H,I), for CN21. (L) Average frequency increase for fast and slow groups, after SCR or CN21. Bootstrap with Bonferroni correction. *p < 0.05. N = 39 neurons (SCR-T1: 11, SCR-T2: 8, CN21-T1: 8, CN21-T2: 12), from 5 rats (1–4 slices per rat).

FIGURE 4

Early CaMKII inhibition abolishes or occludes homeostatic adaptations in different groups of synapses. (A) Overlaid τ_r-sorted mEPSC amplitude distributions for each temporal stage and condition. The curve for CN21 at T2 was omitted for simplicity. (B) Amplitude percent differences with respect to the SCR distribution at T1 (C) Same as panel (A), for decay time. Inset, detail for fast-rising group. (D) Same as panel (B), for decay time. (E) Superimposed event frequency distributions for each time interval and condition (same curves as in Figures 3H,J; left). (F) Frequency increases with respect to SCR condition at T1. Same data as for Figure 3. Bootstrap with weighted p-values (all panels). *p < 0.05.