Perineuronal Nets Enhance the Excitability of Fast-Spiking Neurons

Abstract

Perineuronal nets (PNNs) are specialized complexes of extracellular matrix molecules that surround the somata of fast-spiking neurons throughout the vertebrate brain. PNNs are particularly prevalent throughout the auditory brainstem, which transmits signals with high speed and precision. It is unknown whether PNNs contribute to the fast-spiking ability of the neurons they surround. Whole-cell recordings were made from medial nucleus of the trapezoid body (MNTB) principal neurons in acute brain slices from postnatal day 21 (P21) to P27 mice. PNNs were degraded by incubating slices in chondroitinase ABC (ChABC) and were compared to slices that were treated with a control enzyme (penicillinase). ChABC treatment did not affect the ability of MNTB neurons to fire at up to 1000 Hz when driven by current pulses. However, f-I (frequency-intensity) curves constructed by injecting Gaussian white noise currents superimposed on DC current steps showed that ChABC treatment reduced the gain of spike output. An increase in spike threshold may have contributed to this effect, which is consistent with the observation that spikes in ChABC-treated cells were delayed relative to control-treated cells. In addition, parvalbumin-expressing fast-spiking cortical neurons in >P70 slices that were treated with ChABC also had reduced excitability and gain. The development of PNNs around somata of fast-spiking neurons may be essential for fast and precise sensory transmission and synaptic inhibition in the brain.

Keywords: auditory brainstem; extracellular matrix; fast spiking; inhibition; perineuronal net; white noise current.

Graphical abstract

Figure 1.

ChABC effectively degrades PNNs in acute brain slices during slice recovery. A, WFA-labeled PNNs are visible surrounding MNTB neurons after P-ase treatment (negative control). Scale bars: 200 μm; inset, 20 μm. B, ChABC-treated slices had reduced PNN labeling. C, Maximum pixel intensities in MNTB were reduced by ChABC (t test, p = 0.0007). Error bars are the mean and SEM. D, Cumulative plot of pixel intensities shows that ChABC-treated MNTB regions have reduced brightness (two-sample Kolmogorov–Smirnov, p < 0.008). Slices were labeled in parallel with the same reagents and imaged with the same microscope settings.

Figure 2.

Passive membrane properties were not affected by ChABC treatment. A, Response to current steps show that hyperpolarization activated transient, outward rectification and a single spike on depolarizing steps. Inset shows a spike with an extended time base. Inset calibration: 20 mV, 1 ms. B, ChABC treatment did not affect the I–V curve measured when the membrane reached steady state (as indicated by the open circle in A), or C, size of the hyperpolarizing transient, measured at the peak (filled circle in A). Error bars are the mean and SEM. D, Both P-ase-treated (left) and ChABC-treated (right) neurons were able to fire at 1000 Hz in response to a train of 4 nA, 0.1 ms pulses.

Figure 3.

Spikes evoked by 450 pA, 500 ms current step were not affected by ChABC treatment. Data points are individual cells. Error bars are the mean and SEM.

Figure 4.

ChABC-treated MNTB neurons exhibited lower gain. A, Example white noise currents with noise level of 400 pA SD at four levels of DC step (top), spiking response in a P-ase-treated cell (middle, black), and spiking response in a ChABC-treated cell (bottom, red). B, Expanded time base showing spiking responses from the traces in A at the time indicated by underlining in A in the 600 pA DC step condition. C, f–I curves indicate significantly lower firing rates in ChABC-treated cells compared with P-ase-treated cells (*two-way RM ANOVA, p < 0.05). D, Maximum gain calculated from the f–I curves was significantly lower in ChABC-treated cells, indicating reduced excitability (*two-way RM ANOVA, p < 0.05). E, The distribution of instantaneous firing rates was shifted to lower frequencies and did not indicate a change in spiking pattern.

Figure 5.

Input resistance and resting membrane potential were not different between treatment groups. A, Traces were low-pass filtered to calculate the membrane potential during white noise current stimulation. Thick dark line is the filtered trace plotted over the original trace. Dashed horizontal line indicates the mean membrane potential (Vm) during the current step. These values were used to compare the membrane potential during DC steps shown in B. Examples traces are at the 400 pA noise level and 600 pA DC step. B, There were no significant differences between treatment groups at any noise level (two-way RM ANOVA, p > 0.05). C, Resting membrane potential (Vrest) was calculated as the membrane potential before the current injection. There were no differences between treatment groups at any noise level (two-way RM ANOVA, p > 0.05).

Figure 6.

STAs of injected currents reveal higher current threshold in ChABC-treated MNTB neurons. Injected currents that triggered spikes were aligned and averaged. A, Overlaid examples of P-ase-treated (black) and ChABC-treated (red) cells at the 400 pA SD noise level superimposed on four DC current steps. The peaks of the action potentials were aligned with the rightmost point of each STA. Note that in these examples the peak of the STA is higher in the ChABC-treated cell than in the P-ase-treated cell. B, The average STA peak across cells was significantly higher among the ChABC-treated cells than the P-ase-treated cells (*two-way RM ANOVA, p < 0.05). Error bars are the mean and SEM.

Figure 7.

Voltage threshold and spike shape during fast spiking. A, Phase plane plots of white noise evoked spikes during 400 pA SD noise level without a DC step averaged across ChABC-treated (red) and P-ase-treated (black) cells. The membrane potential where the dV/dt begins to increase is more depolarized in ChABC-treated cells, indicating a more depolarized voltage threshold. Also note that the upstroke of the spike is slower (lower peak dV/dt) and reaches a lower membrane potential. B, Voltage thresholds were significantly depolarized in ChABC-treated cells compared with P-ase-treated cells (see Results). Each marker indicates a DC step (from 0 to 600 pA) plotted at the average evoked firing rate and the average voltage threshold during the step. C, Spike amplitude was significantly smaller in ChABC-treated cells. D, The acceleration of the membrane potential was significantly lower in ChABC-treated cells. Error bars are the mean and SEM.

Figure 8.

ChABC-treated cells consistently fired later than P-ase-treated cells. A, Each overlaid trace is from a different P-ase-treated (black) or ChABC-treated (red) cell. A dot raster is plotted above the spike trains, with each row indicating a different cell. The PSTH below indicates how spike events were identified. When a 0.25 ms bin reached a threshold where ≥25% of P-ase-treated cells fired, a mean spike time was calculated for spikes occurring within 1 ms centered on the bin. The vertical lines above indicate the mean spike time of the spikes within each 1 ms spike event. B, The probability of the ChABC-treated cells to fire later than the P-ase-treated cells during these spike events was high (> 50%) across all levels of gain and DC steps. C, The distribution of the delays of spike events is shifted to the right, indicating that spikes usually occurred later in ChABC-treated cells than P-ase-treated cells.

Figure 9.

Parvalbumin-expressing fast-spiking cortical neurons were similarly affected by ChABC. A, Example of a PV+ GFP-expressing cell that was filled with a fluorophore during recording (Alexa Fluor 594; red) and labeled post hoc with WFA (cyan). Scale bar, 10 um. B, Example traces showing responses to white noise current (400 pA SD, 300 pA DC step) in P-ase-treated cells (black) and ChABC-treated cells (red). C, Fast-spiking cortical neurons treated with ChABC spiked significantly less than P-ase-treated cells in response to white noise currents (*two-way RM ANOVA, p < 0.05). D, Maximum gain calculated from the f–I curves was significantly lower in ChABC-treated cells, indicating reduced excitability (*two-way RM ANOVA, p < 0.05). E, Overlaid traces from each P-ase-treated (black) and ChABC-treated (red) cell, illustrating that ChABC-treated cells fired spikes later than P-ase-treated cells. F, Histogram showing rightward shift in delay, indicating that most spike events from ChABC-treated cells were delayed relative to P-ase-treated cells.