Traumatic Brain Injury Increases Cortical Glutamate Network Activity by Compromising GABAergic Control

Abstract

Traumatic brain injury (TBI) is a major risk factor for developing pharmaco-resistant epilepsy. Although disruptions in brain circuitry are associated with TBI, the precise mechanisms by which brain injury leads to epileptiform network activity is unknown. Using controlled cortical impact (CCI) as a model of TBI, we examined how cortical excitability and glutamatergic signaling was altered following injury. We optically mapped cortical glutamate signaling using FRET-based glutamate biosensors, while simultaneously recording cortical field potentials in acute brain slices 2-4 weeks following CCI. Cortical electrical stimulation evoked polyphasic, epileptiform field potentials and disrupted the input-output relationship in deep layers of CCI-injured cortex. High-speed glutamate biosensor imaging showed that glutamate signaling was significantly increased in the injured cortex. Elevated glutamate responses correlated with epileptiform activity, were highest directly adjacent to the injury, and spread via deep cortical layers. Immunoreactivity for markers of GABAergic interneurons were significantly decreased throughout CCI cortex. Lastly, spontaneous inhibitory postsynaptic current frequency decreased and spontaneous excitatory postsynaptic current increased after CCI injury. Our results suggest that specific cortical neuronal microcircuits may initiate and facilitate the spread of epileptiform activity following TBI. Increased glutamatergic signaling due to loss of GABAergic control may provide a mechanism by which TBI can give rise to post-traumatic epilepsy.

Keywords: cortical hyperexcitability; epilepsy; glutamate signaling; network reorganization; traumatic brain injury.

Figure 1.

Cortical electrophysiological activity is increased following TBI. Representative traces of evoked field potentials from sham- (A) and CCI-injured cortex (B). Inset in (B), outlined by gray box, highlights epileptiform activity occurring during the first 175 ms in CCI-injured cortex. Cortical field potential responses were collected from a recording electrode placed in layer V with a stimulating electrode at the layer VI–white matter interface at an ISI of 30 s. Traces from CCI-injured cortex show increased high-frequency epileptiform activity with increased fEPSP amplitude and duration. (C) Quantification of % epileptiform activity, the number of recordings showing epileptiform activity was divided by total number of recordings per slice, ***P < 0.001, χ² test. Representative evoked field responses at threshold (gray traces) and 2× threshold stimulation (black traces) for sham (D), and CCI-injured cortex (E). (F) Average peak amplitude of field responses showing a significant difference in sham-injured cortex between threshold and 2× threshold stimulation, **P < 0.01, bars represent mean + SEM, pair-sample t-test, n = 16 sham, 12 CCI slices. Five representative cortical field potential recordings from sham (G) and CCI-injured cortex (H) showing after-discharges occurring after initial evoked stimulus. (I) Quantification of % after-discharges, the number of recordings showing after-discharges was divided by total number of recordings per slice, ***P < 0.001, bars represent mean + SEM, two-sample t-test, n = 16 sham, 12 CCI slices.

Figure 2.

Epileptiform activity after TBI is NMDA receptor dependent. Representative cortical field potential recordings from CCI injured slice showing epileptiform activity (A), ∼5 min after perfusion of CPP (B), and DNQX (C). Epileptiform activity is eliminated after blocking NMDA receptors with CPP. Cortical field potential is completely abolished after AMPA receptor blockade with DNQX. (D) Quantification of cortical field potential amplitude showing a significant decrease after CPP and DNQX perfusion compared with aCSF, amplitude from DNQX is significantly decreased compared with CPP. (E) Quantification of cortical field potential area during the first 250 ms after evoked stimulus shows a significant decrease in CPP and DNQX compared with aCSF. (F) Quantification of cortical field potential coastline during the first 250 ms after evoked stimulus shows a significant decrease in CPP and DNQX compared with aCSF; ***P < 0.05 compared with aCSF; ^###P < 0.05 between CPP and DNQX, bars represent mean + SEM, n = 8 slices.

Figure 3.

Glutamate signaling is increased in CCI-injured cortex at threshold stimulation. (A) Representative bright field image of a CCI injured cortical slice demonstrating the massive loss of cortical tissue (left) 2–4 weeks post-TBI. Arrows point to pial surface (dotted line), superficial cortical layers, deep cortical layers, and white matter. Representative, pseudo-colored glutamate images showing the peak glutamate signal (ΔFRET_signal/ΔFRET_noise) for each pixel from sham- (B) and CCI-injured cortex (C). (D) Individual ΔFRET_signal/ΔFRET_noise traces over time for all sham (black) and CCI injured (red) recordings. (E) Averaged ΔFRET_signal/ΔFRET_noise traces over time for all sham (black) and CCI injured (red) recordings. Black arrows indicate time of stimulation. (F) Quantification of peak glutamate signal (ΔFRET_signal/ΔFRET_noise) showing a significant increase in CCI-injured cortex compared with sham. (G) Quantification of activated cortical area, data are expressed as percentage of pixels above 2.5 SDs of the noise, ***P < 0.001 compared with sham, bars represent mean + SEM, two-sample t-test. (H) Correlation of fEPSP amplitude (mV) versus peak ΔFRET_signal/ΔFRET_noise from sham- and CCI-injured cortex (R² = 0.83), n = 75 sham, 50 CCI recordings.

Figure 4.

Disinhibited slices from sham and CCI cortex show similar electrophysiological parameters and glutamate signaling. Representative evoked field responses from sham- (A) and CCI-injured cortex (B) show similar increases in electrophysiological parameters after 10–15 min local perfusion of 10 μM GBZ. Representative, pseudo-colored glutamate signal images showing the peak glutamate signal (ΔFRET_signal/ΔFRET_noise) from sham- (A, inset) and CCI-injured cortex (B, inset) after 10–15 min local perfusion of 10 μM GBZ. (C) Averaged ΔFRET_signal/ΔFRET_noise traces over time for sham (black) and CCI injured (red) recordings. Black arrow indicates time of stimulation. (D) Quantification of peak glutamate signal (ΔFRET_signal/ΔFRET_noise) showing no difference between sham- and CCI-injured cortical slices. (E) Quantification of activated cortical area, data are expressed as percentage of pixels above 2.5 SDs of the noise; bars represent mean + SEM, n = 71 sham, 50 CCI recordings.

Figure 5.

Glutamate signaling is enhanced in proximal deep cortical layers following traumatic brain injury. (A) Illustration showing cortical subregions for glutamate signal analysis, # indicates position of stimulating electrode, * indicates position of recording electrode. Representative, pseudo-colored glutamate images showing the peak glutamate signal (ΔFRET_signal/ΔFRET_noise) from CCI-injured cortex at threshold stimulation (B), CCI-injured cortex in GBZ (D), and sham-injured cortex in GBZ (F). Circles indicate layer-specific cortical subregions relative to the area adjacent to the injury. Quantification of layer-specific peak glutamate signaling in CCI-injured cortex at threshold stimulation (C), proximal deep layers are significantly different from proximal superficial layers and distal deep layers. In disinhibited (10 μM GBZ) CCI-injured cortex (E) and disinhibited sham-injured cortex (G), proximal deep layers are significantly increased compared with distal deep layers; cortical deep layers are significantly increased from superficial layers both proximally and distally, *P < 0.05, **P < 0.01, ***P < 0.001, paired t-test, bars represent mean + SEM, n = 15 sham, 10 CCI slices.

Figure 6.

Temporal analysis of glutamate signaling in the cortex following TBI. (A, C, E) Time-integrated ΔFRET_signal/ΔFRET_noise images from 5-ms binned increments illustrating regional specific onset of glutamate signal in CCI-injured cortex at threshold stimulation (A), disinhibited CCI-injured cortex in 10 μM GBZ (C), and disinhibited sham-injured cortex in 10 μM GBZ (E); post-stimulation time of each frame is indicated in white in upper left corner of each image. (B, D, F) Time of local signal onset from CCI-injured cortex at threshold stimulation (B), disinhibited CCI-injured cortex in 10 μM GBZ (D), and disinhibited sham-injured cortex in 10 μM GBZ (F). Time of signal onset was calculated as the average time in milliseconds at which pixels within a given cortical subregion crossed a threshold of 2.5 SDs above the noise. Superficial and deep areas proximal to the injury activate earlier than superficial and deep distal areas; deep layers are activated earlier than superficial layers both proximally and distally, *P < 0.05, **P < 0.01, ***P < 0.001, paired-t test. (G) Diagram illustrating the sequential time of local signal onset within cortical subregions. (H) Time of global signal onset from whole slices showing CCI-injured cortex in GBZ activates earlier than CCI-injured cortex at threshold stimulation and sham-injured cortex in GBZ, *P < 0.05, two-sample t-test, bars represent mean + SEM, n = 15 sham, 10 CCI slices.

Figure 7.

Number of parvalbumin (+) and somatostatin (+) GABAergic interneurons decrease following TBI. Representative images from PV (A) (red, left) and SST (E) (red, left) from sham- and CCI-injured cortical sections. NeuN images are shown in green and merged images are in yellow. Fewer PV+ and SST+ GABAergic interneurons are seen in CCI- compared with sham-injured cortex. Global quantification of PV+ (B) and SST+ (F) cells per 10 000 μm² cortex from the entire cortical slice shows a significant decrease in cell number in CCI compared with sham-injured cortex, paired t-test. (C) Quantification of PV+ cells per 10 000 μm² in CCI injured deep cortical subregions showing a significant decrease in cell number in the area adjacent to the CCI lesion, proximal deep and distal deep layers compared with deep sham layers. The area adjacent to the injury was significantly decreased compared with proximal and distal deep layers. (D) Quantification of PV+ cells per 10 000 μm² in CCI injured superficial cortical subregions showing a decrease in cell number in the area adjacent to the CCI lesion, proximal superficial and distal superficial layers compared with deep cortical sham layers. The area adjacent to the injury was significantly decreased compared with proximal and distal superficial layers, ***P < 0.001 compared with sham, ^##P < 0.01 compared with proximal and distal regions, one-way ANOVA, bars represent mean + SEM, n = 3 sham, 4 CCI animals; 22 sham, 40 CCI sections. (G) Quantification of SST+ cells per 10 000 μm² in CCI injured deep cortical subregions showing a significant decrease in cell number in the area adjacent to the CCI lesion, proximal deep and distal deep layers compared with deep sham layers. (H) Quantification of SST+ cells per 10 000 μm² in CCI injured superficial cortical subregions showing a decrease in cell number in the area adjacent to the CCI lesion, proximal superficial, and distal superficial layers compared with deep cortical sham layers, **P < 0.01, ***P < 0.001 compared with sham, one-way ANOVA, bars represent mean + SEM, n = 3 animals, 16 sections.

Figure 8.

GLT-1 and GLAST protein expression does not change after TBI. Western blot analysis of cortical lysates prepared from sham- and CCI-injured cortex 2–4 weeks postinjury. Samples were probed for the astrocytic glutamate transporters GLT-1 (A) and GLAST (C). Protein expression was quantified by dividing optical density of GLT-1 (B) or GLAST (D) to β-actin, values are expressed as fold increase relative to sham controls, n = 3 mice.

Figure 9.

TBI increases the frequency of sEPSCs and decreases the frequency of sIPSCs. (A) Representative traces of sEPSCs collected from cortical layer V after sham (black) or CCI injury (gray). (B) Superimposed representative average sEPSCs show no difference in amplitude between sham and CCI slices. Cumulative probability distribution of sEPSC amplitude (C) and average amplitude (D) show no significant difference between sham- and CCI-injured cortex. Cumulative probability distribution of sEPSC interevent interval (E) and average frequency (F) show a significant increase in CCI-injured cortex compared with sham; ***P < 0.001, K-S test; *P < 0.05, mean + SEM, two-sample t-test; n = 9 sham, 11 CCI cells. (G) Representative traces of sIPSCs from sham (black) and CCI (gray) cortex. (H) Superimposed representative average sIPSCs demonstrate no difference in amplitude between sham and CCI cortex. Cumulative probability distribution of sIPSC amplitude (I) and average amplitude (J) show no significant difference between sham and CCI. Cumulative probability distribution of sEPSC interevent interval (K) and average frequency (L) show a significant decrease in CCI-injured cortex compared with sham; ***P < 0.001, K-S test; *P < 0.05, mean + SEM, two-sample t-test; n = 8 sham, 12 CCI cells.