Mice with behavioral evidence of tinnitus exhibit dorsal cochlear nucleus hyperactivity because of decreased GABAergic inhibition

Abstract

Tinnitus has been associated with increased spontaneous and evoked activity, increased neural synchrony, and reorganization of tonotopic maps of auditory nuclei. However, the neurotransmitter systems mediating these changes are poorly understood. Here, we developed an in vitro assay that allows us to evaluate the roles of excitation and inhibition in determining the neural correlates of tinnitus. To measure the magnitude and spatial spread of evoked circuit activity, we used flavoprotein autofluorescence (FA) imaging, a metabolic indicator of neuronal activity. We measured FA responses after electrical stimulation of glutamatergic axons in slices containing the dorsal cochlear nucleus, an auditory brainstem nucleus hypothesized to be crucial in the triggering and modulation of tinnitus. FA imaging in dorsal cochlear nucleus brain slices from mice with behavioral evidence of tinnitus (tinnitus mice) revealed enhanced evoked FA response at the site of stimulation and enhanced spatial propagation of FA response to surrounding sites. Blockers of GABAergic inhibition enhanced FA response to a greater extent in control mice than in tinnitus mice. Blockers of excitation decreased FA response to a similar extent in tinnitus and control mice. These findings indicate that auditory circuits in mice with behavioral evidence of tinnitus respond to stimuli in a more robust and spatially distributed manner because of a decrease in GABAergic inhibition.

Fig. 1.

Example of flavoprotein autofluorescence (FA) signal evoked by electrical stimulation of dorsal cochlear nucleus (DCN) brain slices prepared from control mice (6–13 wk old). (A) An image from a brainstem slice including the DCN. A stimulating electrode was placed in the molecular layer of the DCN to stimulate parallel fibers. (B) Electrical stimulation (100-Hz stimulation for 1 s is indicated by solid bar) resulted in evoked FA signal. FA signals represent percent increases relative to baseline (black). To confirm that these signals were dependent on neuronal action potentials, we applied TTX and observed that electrical stimulation no longer resulted in an increase of fluorescence above baseline (red). Vertical dashed lines illustrate the temporal window that we used to calculate average FA response for all successive analysis. This window begins 0.5 s before electrical stimulation and ends 0.5 s after the termination of electrical stimulation. (C) Single frames of the DCN optical field illustrate the spatial spread and magnitude of the FA response. A time point preceding stimulation and a time point during the peak of the response are shown. Arrowheads indicate the position of the stimulating electrode.

Fig. 2.

Behavioral evidence of tinnitus. (A) Control and tinnitus mice show normal startle reflex (a2) in response to a startle stimulus (a1; 115 dB for 20-ms duration embedded in a 70-dB constant background sound). Startle responses represent the time course of the downward force that the mouse applies onto the platform after the startle pulse (a2 and a4). Previous studies have shown that control animals detect the silent gap (a3) and that their startle reflex is inhibited robustly (a4, black trace). However, animals with behavioral evidence of tinnitus have difficulty detecting the gap when the frequency of the background sound is at or near the frequency of their putative tinnitus. Thus, their startle reflex is less inhibited (a4, gray trace). (B) Noise-exposed mice with behavioral evidence of tinnitus. Relative startle ratio is (response to gap + startle stimulus)/response to startle alone. Background noise was at 24 KHz. A response of one suggests no gap detection (no inhibition of the startle). Noise-induced mice revealed behavioral evidence of tinnitus 2–9 wk after exposure (control: relative startle = 0.55 ± 0.04, n = 8; tinnitus: relative startle = 0.85 ± 0.05, n = 8; P < 0.05). Only mice that showed increased relative startle ratios are included in this graph (about 50% of noise-induced mice). (C) ABR thresholds reveal only temporary threshold shifts for 24-KHz tones. Noise-exposed mice showed significant postexposure threshold elevation (control = 30 ± 5 dB; postexposure = 60 ± 6 dB; P < 0.05). This temporary elevation was recovered by the time that noise-exposed mice were behaviorally assessed (2–9 wk postexposure = 38 ± 7 dB). For the noise-induced mice in C, ABR measurements were obtained from the sound-exposed (ipsilateral) ears. (D) When a 50-ms pulse precedes the startle pulse by 130 ms (left side), the startle responses are similar in control (black) and noise-induced mice (gray). (E) Prepulse inhibition measured after a brief (50 ms) duration stimulus (10 KHz) was similar in controls and mice with behavioral evidence of tinnitus (control PPI = 0.35 ± 0.03; tinnitus PPI = 0.39 ± 0.05). Error bars indicate SEM.

Fig. 3.

FA responses to electrical stimulation are larger and more widespread in the DCN of tinnitus mice. (A) Single-image frame during electrical stimulation of a brain slice from a control mouse illustrating localized FA responses. All successive analysis was based on the average signals obtained from three regions of interest (ROIs; white squares)—the center ROI and the two surrounding ROIs. The site of the stimulation electrode defined the center pixel of the center ROI. Surround ROIs were positioned on either side of the center ROI along the fusiform cell layer. We measured a 200-μm distance from the center pixel of the center ROI. This measurement determined the center pixel of surrounding ROIs. (B) Single-image frame during electrical stimulation of a brain slice from a tinnitus mouse illustrating FA responses that spread to surrounding ROIs. (C) The population average of center responses is larger than the responses recorded at surround regions in control mice (Left; n = 11). The population average of responses in tinnitus mice reveals significantly smaller differences in responses from center and surrounding regions (Right; n = 12). The gray zone indicates SEM. (D) The ratio of the average surround response to the center response is significantly lower in tinnitus mice (control = 0.34 ± 0.04, n = 11; tinnitus = 0.63 ± 0.06, n = 12; P < 0.01). Surround signal for all graphs is the average of the signal obtained from both surrounding ROIs. (E) FA signal propagation to surrounding ROIs in tinnitus mice involves active processes. We used a sharp electrode to bisect the molecular layer near the middle of the DCN (Left, dashed yellow line). A stimulating electrode was placed on one side of the transection. Only sites located on the side of the transection where the stimulating electrode was positioned show FA responses. (F) Similar experiment performed in a different slice. Recordings on the side of the transection where the stimulating electrode is positioned reveal FA response, whereas sites on the other side do not. When the stimulating electrode is repositioned on the other side of the transection, FA response now appears on that side (Right). Error bars indicate SEM.

Fig. 4.

Input–output functions reveal increased responsiveness of DCN FA responses for weak stimuli in tinnitus mice. (A) Population-averaged FA responses (from control mice, n = 11) to stimuli of different durations. All stimuli were composed of 100-Hz pulses. Responses were calculated as the average signal over the duration of the pulse train, including 0.5 s before and after the stimulus. (B) Input–output response functions for center and surrounding signals in control and tinnitus mice. Shaded areas indicate the regions for weak (50–100 pulses) and strong (200–400) stimuli. (C) Comparison of the average slope of input–output functions between control and tinnitus. The slope of the response curves was calculated as the difference in the signal amplitude between two points divided by the difference in the corresponding number of pulses delivered. The slopes of the center response curves for weak stimuli (Left) are similar in control (n = 11) and tinnitus (n = 12) slices. However, the average slope of the surround response curves for weak stimuli is significantly steeper in tinnitus mice (control = 0.40 ± 0.08 × 10⁻²% per pulse, n = 11; tinnitus = 1.16 ± 0.29 × 10⁻²% per pulse, n = 12; P < 0.05). The average slopes of center and surrounding response curves for strong stimuli (Right) are similar in tinnitus and control DCN slices. Error bars indicate SEM.

Fig. 5.

Decreased GABAergic inhibition in tinnitus mice. (A) Individual traces for center signals from control (Left) and tinnitus mice (Right) before and after successive application of strychnine (STR; glycine receptor antagonist = 0.5 μM), SR-95531 (SR; GABA receptor antagonist = 20 μM), NBQX (AMPA receptor antagonist = 10 μM), and AP5 (NMDA receptor antagonist = 100 μM). (B) Individual traces for surround signals from control (Left) and tinnitus mice (Right) before and after successive application of STR, SR, NBQX, and AP5. (C) Average values of the relative center response amplitudes (normalized to no antagonist values) after sequential application of blockers of STR, SR, NBQX, and AP5 reveal no differences between control and tinnitus mice. (D) Average values of the relative surround response amplitudes (normalized to no antagonist values) after sequential application of blockers of STR, SR, NBQX, and AP5 reveal differences in GABAergic inhibition between control and tinnitus mice. There is a significant difference in the relative increase in signal amplitude after SR application (control: normalized ΔF/F₀ after SR = 1.37 ± 0.13, n = 11; tinnitus: normalized ΔF/F₀ after SR = 1.03 ± 0.04, n = 12; P < 0.05). STR, NBQX, and AP5 responses for control and tinnitus mice showed no difference on the relative change. For all experiments, drug application was sequential: STR was added first in the bath, then SR95531, NBQX, and finally, AP5. Error bars indicate SEM.