Abstract
Layer 6 corticothalamic neurons (CTs) provide strong feedback input that is crucial to perception and cognition in normal and pathological states; however, the synaptic properties of this input remain largely unknown, especially in pathology. Here, we examined the synaptic properties of CT axon terminals in the medial geniculate body (MGB), the auditory thalamus, in normal hearing male and female mice and in a mouse model of noise-induced hearing loss (NIHL), also in male and female mice. In normal hearing mice, we found that the amplitude of CT-evoked excitatory postsynaptic current (EPSC) to the core-type ventral subdivision of the auditory thalamus (MGv), which mainly conveys rapid sensory information, is larger compared with the amplitude of CT-evoked EPSC to the matrix-type dorsal subdivision of the auditory thalamus (MGd), which likely conveys higher-order internal state information. This is due to higher axonal density and/or axonal recruitment in CT→MGv compared with CT→MGd synapses. After noise trauma, we observed enhanced short-term facilitation in CT→MGd but not CT→MGv synapses. Our findings reveal a previously unknown mechanism of short-term synaptic plasticity after NIHL via which CTs enhance the throughput of matrix-type thalamus, likely to improve perceptual recovery via higher-order contextual modulation.
Significance Statement
Auditory layer 6 corticothalamic neurons (CTs) send massive projections to the auditory thalamus, the medial geniculate body (MGB). This pathway is crucial for sound perception and cognition. However, the synaptic properties of this pathway under either normal or pathological hearing remain poorly understood. We found enhanced evoked synaptic responses between CT and the ventral (CT→MGv) compared with CT and the dorsal (CT→MGd) MGB subdivision. Importantly, we discovered an enhancement of activity-dependent facilitation at the CT→MGd synapses after noise-induced hearing loss, thus highlighting a plasticity mechanism that might enhance perceptual recovery via higher-order cortico-thalamo-cortical modulation.
Introduction
Cortico-thalamo-cortical circuits are critically involved in perception and consciousness (Guo et al., 2017; Homma et al., 2017; Voigts et al., 2020; Shepherd and Yamawaki, 2021) and thus provide a crucial network for studying synaptic mechanisms underlying perception and recovery of perception after sensory organ damage. In the cortico-thalamo-cortical loop, the reciprocity of projections between cortical and thalamic areas allows for two or more brain regions to concurrently stimulate and be stimulated by each other. This reciprocal, hierarchical, and parallel processing (Ji et al., 2016) supports the synchronization and the constant bottom-up and top-down processing that allows for perception and conscious processing of sensory stimuli, such as an auditory or visual scene (Tononi et al., 1998; Tononi and Edelman, 1998; Jones, 2001; Llinas et al., 2005; Sherman, 2007; Edelman and Gally, 2013; Sherman, 2016; Shepherd and Yamawaki, 2021). Cortical projections to the thalamus are mediated by two main different classes of cortical neurons: layer (L) 6 corticothalamic (CT) neurons and L5b pyramidal tract neurons. Here, we will investigate the synaptic organization and properties between CTs and thalamic neurons and the plasticity mechanisms that alter these properties after noise-induced hearing loss (NIHL).
In the visual, somatosensory, and auditory cortex (AC), CT projections to the thalamus adjust the gain of the upper cortical layers (Olsen et al., 2012), as well as the tuning precision, synchrony, and gating of thalamic cells, as required by ongoing behavioral and coding demands (Mease et al., 2014; Guo et al., 2017; Clayton et al., 2021; Ibrahim et al., 2021). As such, CTs can shape both the overall gain and the temporal dynamics of cortical sensory responses (Mease et al., 2014). In the AC, activation of CTs is crucial in gating AC population activity by desynchronizing neurons in the medial geniculate body (auditory thalamus, MGB; Ibrahim et al., 2021) and biasing sound perception toward enhanced detection at the expense of discrimination or vice versa (Guo et al., 2017). Auditory cortex CT neurons project exclusively to the MGB, but the differences in innervation and synaptic properties between CTs and the subdivisions of the MGB are not fully understood. Moreover, how these properties change after noise-induced cochlear damage remains completely unknown.
Given the importance of CT→MGB projections in auditory perception and the prevalence of perceptual deficits NIHL, it is important to understand the plasticity of CT→MGB synapses after noise trauma. Following cochlear damage, despite reduced peripheral auditory input to the brain, the neuronal activity of AC principal neurons is recovered or even enhanced due to compensatory plasticity mechanisms (Qiu et al., 2000; Auerbach et al., 2014; Chambers et al., 2016; Kumar et al., 2023). This plasticity is associated with a remarkable recovery in perceptual thresholds, but other aspects of more complex auditory processing are not compensated (Caras and Sanes, 2015; Chambers et al., 2016; Guo et al., 2017; Kumar et al., 2023). Moreover, aberrant forms of this compensatory plasticity can be associated with tinnitus and hyperacusis (Auerbach et al., 2014; Henton et al., 2023). Therefore, understanding the underlying mechanisms of plasticity following loss of peripheral input is crucial for highlighting strategies that might mitigate hearing loss and hearing loss-related disorders.
Although recent studies on AC layer (L) 2/3 revealed a cell-type specific synaptic, intrinsic and circuit plasticity among principal neurons (Kumar et al., 2023), the plasticity mechanisms in deeper layer neurons, such as the CTs, and their synapses remain largely unknown. To study the precise synaptic mechanisms in the auditory CT→MGB pathway in normal hearing and hearing loss, we used a combination of patch-clamp and anatomical techniques in a NIHL model in transgenic mice (Ntsr1-Cre). In these mice, we can selectively target and activate CT axon terminals and thus explore the potential differences in specific CT synapses. Because the two main subregions of the MGB, the matrix-type (calbindin-expressing) and the core type (Jones, 1998, 2001; Llinas et al., 2005; Larkum, 2013), have distinct projections and roles in sensory processing (Guo et al., 2017; Homma et al., 2017; Voigts et al., 2020; Shepherd and Yamawaki, 2021), we focused our studies on comparing the properties of CT synapses between these subdivisions of the MGB in normal and pathological hearing. Our results reveal MGB subregion-specific differences in the baseline synaptic transmission in normal hearing and in short-term plasticity (STP) in the CT→MGB pathway after hearing loss, which can dynamically gate selective thalamic throughput, likely to meet the challenging perceptual demands after hearing loss.
Materials and Methods
Animals
We used 35 male and 34 female Ntsr1-Cre mice [GENSAT, B6.FVB(Cg)-Tg(Ntsr1-Cre)GN220Gsat/Mmcd] for all experiments. All mice were handled in concordance with the National Institute of Health guidelines and approved by the Institutional Animal Care and Use Committee at the University of Pittsburgh. Experiments and analyses were blinded to the noise or sham exposure condition.
Stereotaxic injections
Ntsr1-Cre mice [postnatal days (P) 28–35] were anesthetized with inhaled isoflurane (induction, 3% in oxygen; maintenance, 1.5% in oxygen) and secured in a stereotaxic frame (Kopf). Core body temperature was maintained at ∼37°C with a heating pad, and eyes were protected with ophthalmic ointment. A small craniotomy was made for intracranial injections in the right AC [lambda (mm); a/p: 0.0, m/l: 4.0, d/v: −1.0, with needle angled 30°], with a micromanipulator (Kopf).
Given that MGB neurons also project to amygdala (Doron and Ledoux, 1999, 2000), and because we focused our studies in cortico-thalamo-cortical loops, we injected ∼0.1 µl red retrograde fluorescence latex microspheres (retrobeads, Lumafluor) in the right AC to label and subsequently recorded from thalamocortical neurons in the MGB. The retrobeads were pressure injected (25 psi, 10–15 ms duration) from capillary pipettes (Drummond Scientific) with a Picospritzer (Parker Hannifin). To express ChR2 in the L6 corticothalamic pathway, we injected AAV9-EF1a-double floxed-hChR2-EYFP (titer: 2.1e13 vg/mL, 1:10 dilution in PBS, Addgene, catalog #20298) in the right AC over the course of 3 min, using a syringe pump (World Precision Instruments). The pipette was maintained in place for 20 min and then withdrawn carefully.
Brain slice electrophysiological recordings
Patch-clamp whole-cell recordings were performed 3–4 weeks after AAV viruses and retrobead injections. Mice (P50–70) were first anesthetized with isoflurane and then immediately decapitated. Brains were rapidly removed and coronal slices (300 μm) containing the right AC and MGB were prepared in a cutting solution at 1°C using a vibratome (VT1200 S; Leica). The cutting solution contained the following (in mM): 2.5 KCl, 1.25 NaH2PO4, 25 NaHCO3, 0.5 CaCl2, 7 MgCl2, 7 glucose, 205 sucrose, 1.3 ascorbic acid, and 3 sodium pyruvate (pH 7.4, ∼300 mOsm, bubbled with 95% O2/5% CO2). The slices were immediately transferred and incubated at 34°C in a holding chamber for 40 min before recording. Slices were then stored at room temperature prior to recording. The incubation and recording ACSF contains the following (in mM): 125 NaCl, 2.5 KCl, 26.25 NaHCO3, 2 CaCl2, 1 MgCl2, 10 glucose, 1.3 ascorbic acid, and 3 sodium pyruvate (pH 7.4, ∼300 mOsm, bubbled with 95% O2/5% CO2). The recording solution was maintained at 34°C with an inline heating system (Warner TC324-B). Whole-cell recordings in voltage- and current-clamp mode were performed using a MultiClamp 700B amplifier equipped with Digidata 1440A A/D converter and Clampex (Molecular Devices). Data were sampled at 10 kHz and Bessel filtered at 4 kHz using MultiClamp 700B Commander. For voltage-clamp recordings, borosilicate pipettes (3–5 MΩ, World Precision Instruments) were filled with cesium-based internal solution, containing the following (in mM): 126 CsCH3O3S, 4 MgCl2 10 HEPES, 4 Na2ATP, 0.3 Tris-GTP, 10 Tris-phosphocreatine, and 3 sodium ascorbate (pH ∼7.4, Osmolarity ∼300 mOsm). For current-clamp recordings, pipettes were filled with a potassium-based intracellular solution containing the following (in mM): 113 K-gluconate, 9 HEPES, 4.5 MgCl2, 4 Na-ATP, 0.3 Tris-GTP, 14 Tris-phosphocreatine, 0.1 EGTA. Pipette capacitance was compensated. Series and input resistance were determined by giving a −5 mV hyperpolarizing voltage step for 50 ms in voltage-clamp mode from holding potential at −70 mV, not corrected for junction potential, and were monitored throughout the experiments. Neurons with either >20% of change in series or input resistance throughout the experiment or with series resistance >25 MΩ were excluded from data analysis. The images in Figures 1C–E and 2B were captured with Qcapture software and were processed with ImageJ and Adobe illustrator.
Optogenetic stimulation and strontium quantal events (Sr2+-mEPSCs)
CT terminals that project to MGB were evoked by optogenetic stimulation, by a blue LED light source (470 nm, Thorlabs) through a 40× microscope objective lens. Recordings were targeted to the red retrobeads labeled MGB neurons. Baseline light-evoked excitatory postsynaptic currents (EPSCs) were evoked at 0.1 Hz stimulation frequency with a 10 ms light pulse duration. We initially tried to use the light intensity required to elicit a stable, maximal plateau response, but for most recordings we could not reach a plateau response even at maximal light stimulation. Therefore, for all voltage-clamp recordings, except for minimal stimulation recordings, we used maximal light stimulation. For minimal stimulation recordings, the light intensity stimulation was adjusted to produce a response failure rate >40%. At this intensity, 100 trials were collected per cell with 0.1 Hz stimulation frequency. The average amplitude of minimal stimulation responses (minimal EPSCs) was calculated. STP was obtained by delivering a train of 10 stimuli at 1, 5, and 10 Hz with an interval of 20 s between trains. For current-clamp recordings (Fig. 8), we sometimes had to avoid maximal stimulation intensity to prevent spiking during the train. Paired pulse ratio (PPR) was calculated from the first two pulses of the 10 Hz train. Strontium quantal events were recorded using a modified Sr2+-ACSF solution, which contained the following (in mM): 125 NaCl, 2.5 KCl, 26 NaHCO3, 4 SrCl2, 4 MgCl2, 15 glucose, 1.3 ascorbic acid, and 3 sodium pyruvate, pH 7.4, ∼300 mOsm, oxygenated w/ 95% O2–5% CO2. Slices were incubated in this solution for 30 min prior to recording. All the other conditions were kept the same as described earlier. The same cutting solution, Cs+-based intracellular solution, and the same light activation were used. Quantal events were analyzed (Clampfit 11.2) from a 400 ms window beginning 100 ms after light stimulus (Oliet et al., 1996; Kouvaros et al., 2020). Event detection was optimized by preprocessing the traces with a low-pass Gaussian filter (1 kHz). The amplitude of LED-evoked quantal events was calculated using the equation:
Noise-induced hearing loss
Mice (P40–60) were placed in a customized 5 × 4 box in a sound-isolated acoustic chamber for noise or sham exposure (NE or SE). The noise-exposed mice were exposed bilaterally to an octave band (8–16 kHz) noise at 100 dB SPL for 2 h. Sham-exposed mice underwent the same exposure protocol, but without the presence of noise.
Auditory brainstem responses
Auditory brainstem responses (ABRs) from littermate mice (P40–70) are measured from the left ear, right before SE or NE, and then measured again 1d or 10d after SE or NE, immediately followed by electrophysiological recordings. Mice were placed on a heating pad (∼37°C) in a sound-attenuating chamber (ENV-022SD; Med Associates) and anesthetized under isoflurane (3% induction/1.5% maintenance, in oxygen). ABRs were taken by placing subdermal electrodes at the vertex of the skull (active), under the left ear (reference), and under the right ear (ground). We recorded ABRs by presenting broadband clicks (1 ms duration, 0–80 dB SPL in 10 dB steps) and tones (0–80 dB SPL in 10 dB steps; 10, 12, 16, 20, 24, 32 kHz frequencies; Kouvaros et al., 2020; Marinos et al., 2021). The sound clicks and tones were delivered through a plastic tip placed in the left ear canal at a rate of 18.56 per second with a MF1 speaker (Tucker-Davis Technologies). The speaker was calibrated with a 0.25 inch microphone (4954-B, Brüel & Kjær) using a 1 kHz, 94 dB sound calibrator standard (Type 4231, Brüel & Kjær) as described previously (Marinos et al., 2021). For each stimulus, the evoked response was averaged from 512 trials with bandpass filtering the waveform between 300 and 3,000 Hz. Data acquisition and analyses were performed using Rz6 processor and BioSigRP software. ABR threshold was defined as the lowest stimulus intensity that generated a wave I response.
Tissue preparation for immunohistochemistry
Following electrophysiological recordings, brain slices were postfixed overnight at 4°C in 4% paraformaldehyde in 1× phosphate-buffered saline (PBS), pH 7.4. Slices were washed three times for 10 min in 1× PBS and then either entered immunohistochemistry immediately or placed in 1× PBS + 0.01% sodium azide at 4°C for storage.
Immunostaining
Slices were permeabilized in 1× PBS containing 0.3% Triton X-100 (PBST) three times for 20 min each and then blocked in PBS + 5% normal goat serum (blocking buffer) at room temperature for 2 h. Slices were then incubated in blocking buffer containing primary antibodies against calbindin (1:1,000, Mouse IgG1 anti-calbindin, Swant CB300) for 72 h at 4°C. Following primary antibody incubation, slices were washed three times in PBST for 20 min each. Slices were then incubated in blocking buffer containing the secondary antibody (1:1,000, Goat anti-mouse IgG1 Alexa Fluor 647, Invitrogen A-21240) overnight at 4°C. Following secondary antibody incubation, slices were washed three times in PBST for 20 min each and three times in 1× PBS for 10 min each. Slices were mounted onto slides and coverslipped (ProLong Gold Antifade Mounting Medium).
Confocal image acquisition and analysis
Confocal fluorescent imaging was performed on a Leica Stellaris 5 confocal microscope using a 20× objective and using the program LASX (Leica Microsystems). The approximate area of the slice containing the MGB was identified using the anti-calbindin staining. All images were acquired using the same acquisition settings. Stitched composite images were generated by capturing 2–4 z-stacks (0.5 mm steps, 1× zoom; pinhole, 1.0 Airy unit) of the MGB to capture the entire z-axis of the slice. Fields of view with 10% x–y overlap were captured to encompass the area of interest. The subdivisions of the MGB were determined using histological landmarks from the Allen Mouse Brain Atlas (http://mouse.brain-map.org/static/atlas) and anti-calbindin staining. The ventral portion of the MGv was distinguished from the MGd and MGm based on the anti-calbindin staining, as calbindin labels neurons that are located primarily in the dorsal and medial portions of the MGB (Lu et al., 2009; Kouvaros et al., 2023). In addition to the anti-calbindin staining, the boundaries of MGm were further estimated based on the location of the rhinal fissure, the brachium of the superior colliculus, and the shape of the hippocampal CA3 and dentate gyrus. Regions of interest (ROIs) were drawn in LASX to delineate the MGv, MGd, and MGm. To measure ChR2-eYFP fluorescence intensity, the fluorescence intensity analysis function of LASX was used in each ROI. For comparisons between SE and NE mice, average fluorescence values were normalized to the SE group. To normalize the fluorescence values, for each MGB subdivision, the average SE intensity was calculated, and each individual slice was normalized against this average value, such that the SE group average was set to 1.0. To evaluate the spread of the viral injections in the AC (Fig. 1G,H), we imaged ChR2 fluorescence in the entire hemisphere with a 10× objective at 0.75× digital zoom. We then stitched the different fields of view together (10% x–y overlap; Fig. 1F). We imaged the primary and nonprimary auditory cortices with a 20× objective at 1.0× zoom (10% x–y overlap; Fig. 1G). Auditory cortex boundaries were estimated based on the location of the rhinal fissure and the hippocampus.
Statistics
Prism 9 GraphPad was used for statistical analysis. For statistical comparisons between two independent groups that passed the Shapiro–Wilk normality test, we used unpaired t tests. The Mann–Whitney test was used for non-normally distributed data. For comparisons between multiple groups, one-way, two-way, or two-way ANOVA with repeated measures were used, with post hoc Bonferroni’s correction. Significance levels are denoted as *p < 0.05, **p < 0.01, ***p < 0.001. For detailed values and statistical tests for all figures, see Tables 1⇓–3. Group data are presented as mean ± SEM.
Statistical values for Figure 4
Results
To study the synaptic properties and organization of layer (L) 6 auditory corticothalamic (CT) inputs onto MGB neurons (CT→MGB), we used Ntsr1-Cre transgenic mice (Fig. 1A), which selectively express Cre recombinase in CTs (Fig. 1A, CT neurons, green triangle). Previous studies have confirmed that all Ntsr1+ neurons in the AC are in L6 and are CTs (Guo et al., 2017). To evaluate the CT→MGB synaptic properties and organization, we selectively expressed channelrhodopsin (ChR2) in CTs (Fig. 1A,D, CT axons, green line), by injecting Cre-dependent ChR2 AAV viral vector into the AC. To evaluate the spread of our viral injections, we imaged ChR2 fluorescence, which showed reliable ChR2 expression throughout the primary and nonprimary auditory cortical fields (Fig. 1F,G). To label MGB neurons that project to the AC (Fig. 1A,E, thalamocortical MGB neurons red), we injected red microsphere retrobeads into the AC. We then performed whole-cell recordings from acute MGB-containing brain slices (Fig. 1C). We patched onto thalamocortical MGB neurons and recorded light-evoked EPSCs in response to photostimulating CT terminals that expressed ChR2. To evaluate the localization of the MGB neurons in the different MGB subdivisions, we used an anti-calbindin antibody, which preferentially labels the medial (MGm) and dorsal (MGd) but not the ventral (MGv) MGB (Lu et al., 2009; Kouvaros et al., 2023; Fig. 1B). Due to the smaller area, neuronal visibility, and borderline localization of the MGm in our acute brain slices, we were unable to obtain enough whole-cell recordings from that area. We therefore focused our electrophysiological studies on MGv and MGd neurons.
Experimental approach and MGB subdivision identification. A, Left, Schematic illustration of stereotaxic injections of retrobeads to label thalamocortical neurons (MGB neurons, red) and viral vectors (AAVs) for expression of ChR2 in corticothalamic (L6 CT) inputs in the MGB (green). All in Ntsr1-Cre mice. Right, Schematic illustration of whole-cell recordings, including CT ChR2-expressing terminals projecting to MGB principal neurons (thalamocortical neurons), which were labeled retrogradely with red retrobeads injection in AC. Photo-evoked EPSCs were recorded from MGB neurons in response to blue light stimulation. B, Representative MGB image of coronal brain slice from Ntsr1-Cre mice showing distinct MGB subdivisions based on calbindin immunostaining (green). The dorsal (MGd) and medial (MGm) MGB are strongly immunoreactive to calbindin staining. The ventral (MGv) MGB shows little immunoreactivity to calbindin staining. D, dorsal; L, lateral. C, Representative image of an acute coronal brain slice that contains MGd, MGv, and MGm in bright-field (4×). Scale bar is the same as in B. D, Representative MGB image (4×) of coronal section illustrating ChR2-expressing CT projections in the MGB. Scale bar is the same as in B. E, Representative image (4×) red labeled MGB neurons, by retrobeads injected in the AC. Scale bar is the same as in B. F, Representative image from coronal sections illustrating ChR2 expression throughout the hemisphere. Hippocampus (Hippo), rhinal fissure (RF). G, Representative image from coronal sections illustrating ChR2 expression throughout the AC.
CT-evoked EPSCs are stronger in CT→MGv compared with CT→MGd synapses, due to higher axonal density and/or axonal recruitment
We found that the average EPSC peak amplitude was significantly smaller in the CT→MGd synapses compared with CT→MGv synapses (Fig. 2A). Whereas the observed changes in synaptic strength based on optogenetics approaches are subject to potential variable infection efficiency per mouse/brain area, in Figure 2B we show a representative example of recordings from different subdivisions within the same slice, further supporting the effect we observed by averaging EPSCs from MGB neurons from different mice and brain slices.
CT-evoked EPSCs are stronger in CT→MGv compared with CT→MGd synapses, due to higher axonal density and/or axonal recruitment. A, Average light-evoked EPSC amplitude in CT→MGd and CT→MGv synapses (left) and representative traces (right). p = 0.0002 (Mann–Whitney test, CT→MGd: 20 cells; CT→MGv: 30 cells). Data for CT→MGd and CT→MGv are pooled from the 1d SE and 10d SE of CT→MGd and CT→MGv (Fig. 4) recordings, respectively. B, CT→MGB EPSC traces from the same slice. Recorded neurons are numbered from 1 to 5, in order from MGd to MGv. C, Representative average traces of light-evoked quantal EPSCs (Sr2+-mEPSCs; left top) and average Sr2+-mEPSCs amplitude in CT→MGd and CT→MGv synapses (left bottom). p = 0.4299 (unpaired t test, CT→MGd: 15 cells; CT→MGv: 16 cells). Data for CT→MGd and CT→MGv are pooled from the 1d SE and 10d SE of CT→MGd and CT→MGv (Fig. 4) recordings, respectively. D, Representative Sr2+-mEPSCs traces. Left, The top trace (black) is a control trace from Ca2+-containing ACSF; the middle and bottom traces are from Sr2+-containing ACSF. The arrowhead indicates the onset of light stimulus. The solid line indicates the 400 ms time window beginning 100 ms after light stimulus that was used to analyze the amplitude of Sr2+-mEPSC. Right, Amplitude and rise time histograms of events before (background noise, pre-LED) and after (post-LED) stimulation from CT→MGd and CT→MGv synapses are shown. E, Average paired pulse ratio (PPR) in CT→MGd and CT→MGv synapses (left) and representative traces (right). p = 0.4006 (Mann–Whitney test; CT→MGd: 18 cells; CT→MGv: 15 cells). Data for CT→MGd and CT→MGv are pooled from the 1d SE and 10d SE of CT→MGd and CT→MGv recordings (Fig. 4), respectively. F, Average 1/CV2 in CT→MGd and CT→MGv synapses. p = 0.0324 (Mann–Whitney test, CT→MGd: 20 cells; CT→MGv: 30 cells). Data for CT→MGd and CT→MGv are pooled from the 1d SE and 10d SE of CT→MGd and CT→MGv recordings (Fig. 4), respectively. G, Average minimal EPSC amplitude (left) and representative averaged minimal EPSC traces from one cell (right) from CT→MGd (gray) and CT→MGv (black) synapses. p = 0.7567 (unpaired t test, CT→MGd: 5 cells; CT→MGv: 6 cells). H, Representative image of coronal brain slice showing calbindin immunostaining (magenta) and ChR2 fluorescence (green). I, Quantification of ChR2 immunofluorescence intensity in the different MGB subdivisions. p = 0.0027 (one-way ANOVA, MGd: 10 slices; MGv: 10 slices; MGm: 10 slices). Data for MGd and MGv are pooled from the 1d SE and 10d SE of MGd and MGv (Fig. 5C,D,H,I), respectively. Detailed statistical values are listed in Table 1.
To investigate the mechanisms underlying the difference in synaptic strength between CT→MGd and CT→MGv synapses, we employed various assays to evaluate postsynaptic and/or presynaptic mechanisms. First, we collected and analyzed the evoked quantal events in Sr2+ (Sr2+-mEPSCs; Oliet et al., 1996; Kouvaros et al., 2020). Replacing Ca2+ with Sr2+ in the ACSF (Materials and Methods) desynchronizes the evoked neurotransmitter release, thus allowing the analysis of quantal events from the stimulated synapses (Oliet et al., 1996). We found that the average amplitude of the quantal events (q) was not different between CT→MGv and CT→MGd synapses (Fig. 2C,D), supporting that the observed differences in EPSC size are not associated with changes in q.
Because we did not detect any changes in q, we next investigated whether presynaptic mechanisms might account for the differential EPSC amplitude between CT→MGd and CT→MGv synapses. Presynaptic mechanisms were evaluated by analyzing the PPR and 1/CV2 (Faber and Korn, 1991; Zucker and Regehr, 2002). An increase in PPR would indicate a decrease in the probability of neurotransmitter release (p), whereas a decrease in PPR would indicate an increase in p. An increase in 1/CV2 would indicate an increase in the number of functional releasing sites (n) and/or an increase in p, whereas a decrease in 1/CV2 would indicate a decrease in n and/or p. We found that both synapses showed facilitation and PPR was not different between CT→MGd and CT→MGv synapses (Fig. 2E), indicating that p is not different in these synapses. We found a significantly greater 1/CV2 in CT→MGv compared with CT→MGd synapses (Fig. 2F), likely indicating a higher n in L6CT→MGv. Because n is the product of individual synapse strength and number of synapses activated during the stimulus, we used minimal stimulation to determine the synaptic strength after activation of a single axon. We found that minimal response was not different between CT→MGv compared with CT→MGd synapses, suggesting that synaptic strength in response to a single axon stimulation is not different between CT→MGv and CT→MGd synapses (Fig. 2G). Finally, we quantified the average ChR2 immunofluorescence intensity in the MGd, MGv, and MGm nuclei. We found that the average immunofluorescence intensity was significantly higher in MGv than MGd and MGm (Fig. 2H,I). This difference in fluorescence intensity is consistent with a higher density of CT→MGv axons than CT→MGd axons; however, this experiment does not address whether the fluorescence signal reflects axon numbers or axon widths or size of axonal arborizations, nor how many of those axons are just passing through without making any synapses. Taken together, these results support that although neither p, q, or the strength of individual CT inputs are different between CT→MGv and CT→MGd synapses, a higher CT axonal density and/or axonal recruitment in CT→MGv synapses likely underlies the increased EPSC amplitude in CT→MGv compared with CT→MGd synapses.
No changes in baseline synaptic transmission in either CT→MGv or CT→MGd synapses after NIHL
Next, we assessed whether noise trauma affects synaptic properties in either CT→MGv and/or CT→MGd synapses. To answer this question, we used an NIHL mouse model, where we exposed mice bilaterally to an octave band (8–16 kHz) noise at 100 dB SPL for 2 h (Fig. 3A; Materials and Methods; Kumar et al., 2023). As control, we used sham-exposed (SE) mice, which were treated identically to the noise-exposed (NE) mice but without noise presentation (Materials and Methods). To evaluate hearing thresholds, we measured auditory brainstem responses (ABRs), which represent synchronized neural activity in response to sound stimuli from the auditory nerve to the inferior colliculus along the auditory brainstem. The first wave originates from the type I auditory nerve fibers and reflects hearing thresholds (Buchwald and Huang, 1975; Kiang et al., 1976). ABRs were collected before, 1 day (d) after, and 10d after NE or SE (Fig. 3B). ABR threshold shifts were calculated by subtracting the pre-exposure ABR threshold from the postexposure ABR threshold. A positive ABR threshold shift indicates hearing loss, as evidenced by an increase in hearing thresholds after noise exposure. We found that ABR thresholds were elevated at 1d and remained elevated at 10d after NE (Fig. 3C–E). In contrast, ABR thresholds remained unchanged in the 1d or 10d in SE mice (Fig. 3C–E), supporting that our mouse model captures hearing loss that lasts at least for 10 d.
Mouse model of NIHL. A, Noise exposure paradigm. Mice were bilaterally exposed to either an 8–16 kHz sound at 100 dB for 2 h or sham exposure. B, Timeline of experimental design. C, Average ABR threshold shifts (clicks and pure tones) from 1d SE and 1d NE mice. Clicks, p = 0.0019 (Mann–Whitney test, 1d SE: 14 mice; 1d NE: 17 mice); pure tones, p < 0.0001 (effect of exposure, 2-way ANOVA, 1d SE: 15 mice; 1d NE: 17 mice). D, Average ABR threshold shift (clicks and pure tones) from 10d SE and 10d NE mice. Clicks, p = 0.0001 (Mann–Whitney test, 10d SE: 13 mice; 10d NE: 12 mice); pure tones, p < 0.0001 (effect of exposure, 2-way ANOVA, 10d SE: 13 mice; 10d NE: 12 mice). E, Representative ABR traces in response to clicks, from 1d SE, 1d NE, 10d SE, and 10d NE mice. Highlighted traces indicate the ABR thresholds. Detailed statistical values are listed in Tables 1 and 2.
We used this NIHL mouse model to examine the potential changes in the properties of CT→MGv and CT→MGd synapses. Whole-cell recordings of MGB neurons were performed from acutely prepared brain slices 1d or 10d after NE or SE. To study the synaptic properties of the CT→MGB pathway after hearing loss, we patched onto MGB neurons located at either MGd or MGv and recorded light-evoked EPSCs by photostimulating CT terminals that expressed ChR2, as described previously. We found no changes in the EPSC amplitude 1d or 10d after NE in either CT→MGd or CT→MGv synapses (Fig. 4A,E). We did not observe any changes in quantal size (Fig. 4B,F) and 1/CV2 (Fig. 4C,G) between NE and SE in either CT→MGd or CT→MGv synapses. However, we found an increase in PPR 1d after NE in CT→MGd synapses (Fig. 4D), suggesting a decrease in p, which recovers by 10d after NE (Fig. 4H).
No changes in baseline synaptic transmission in either CT→MGv or CT→MGd synapses after NIHL. A, Average ESPC peak amplitude in 1d SE and NE in CT→MGd and CT→MGv synapses (left) and representative traces (right), p = 0.2898 (effect of exposure, 2-way ANOVA, MGd 1d SE: 12 cells; MGd 1d NE: 14 cells; MGv 1d SE: 15 cells; MGv 1d NE: 10 cells). B, Average Sr2+-mEPSC amplitude in 1d SE versus 1d NE in CT→MGd and CT→MGv synapses (bottom left) and representative average traces (top left). The arrowhead indicates the onset of light stimulus. The solid line indicates the 400 ms time window beginning 100 ms after light stimulus that was used to analyze the amplitude of Sr2+-mEPSCs. p = 0.9981 (effect of exposure, 2-way ANOVA, MGd 1d SE: 11 cells; MGd 1d NE: 8 cells; MGv 1d SE: 10 cells; MGv 1d NE: 8 cells). C, Average 1/CV2 in 1d SE versus 1d NE in CT→MGd and CT→MGv synapses. p = 0.5186 (effect of exposure, 2-way ANOVA, MGd 1d SE: 12 cells; MGd 1d NE: 14 cells; MGv 1d SE: 15 cells; MGv 1d NE: 10 cells). D, Average PPR in 1d SE versus 1d NE in CT→MGd and CT→MGv synapses (left) and representative traces (right). p = 0.0193 (effect of exposure, 2-way ANOVA, MGd 1d SE: 9 cells; MGd 1d NE: 7 cells; MGv 1d SE: 7 cells; MGv 1d NE: 11 cells). E, Average ESPC peak amplitude in 10d SE and 10d NE in CT→MGd and CT→MGv synapses (left), and representative traces (right). p = 0.5133 (effect of exposure, 2-way ANOVA, MGd 10d SE: 8 cells; MGd 10d NE: 8 cells; MGv 10d SE: 15 cells; MGv 10d NE: 9 cells). F, Average Sr2+-mEPSC amplitude in 10d SE and 10d NE in CT→MGd and CT→MGv synapses (bottom left) and representative average traces (top left). The arrowhead indicates the onset of the light stimulus. The solid line indicates the 400 ms time window beginning 100 ms after the light stimulus that was used to analyze the amplitude of Sr2+-mEPSCs. p = 0.6887 (effect of exposure, 2-way ANOVA, MGd 10d SE: 4 cells; MGd 10d NE: 5 cells; MGv 10d SE: 6 cells; MGv 10d NE: 6 cells). G, Average 1/CV2 in 10d SE and 10d NE in CT→MGd and CT→MGv synapses. p = 0.5502 (effect of exposure, 2-way ANOVA, MGd 10d SE: 8 cells; MGd 10d NE: 8 cells; MGv 10d SE: 15 cells; MGv 10d NE: 9 cells). H, Average PPR in 10d SE and 10d NE in CT→MGd and CT→MGv synapses (left) and representative traces (right). p = 0.1727 (effect of exposure, 2-way ANOVA, MGd 10d SE: 9 cells; MGd 10d NE: 8 cells; MGv 10d SE: 8 cells; MGv 10d NE: 12 cells). Detailed statistical values are listed in Table 3.
Finally, when we quantified the ChR2 immunofluorescence intensity in MGv and MGd, we did not observe any differences between NE and SE (Fig. 5), consistent with our electrophysiological results. Together, these results support that baseline synaptic transmission at low frequency (0.1 Hz) of CT activation is not altered in either CT→MGv or CT→MGd synapses after NIHL. However, we observed a change in the PPR of CT→MGd synapses when two pulses were delivered at a 10 Hz frequency (100 ms interstimulus interval). This prompted us to investigate whether there are any differences in STP in response to different stimulation frequencies before and after noise trauma.
Unchanged ChR2 fluorescence intensity of CT terminals in the different MGB subdivisions after NIHL. A, B, Representative images of coronal brain slices showing calbindin immunostaining (magenta) and ChR2 fluorescence (green) from 1d SE (A) and 1 NE (B). C–E, Average ChR2 immunofluorescence intensity in the different MGB subdivisions in 1d SE and 1d NE (C, MGd; D, MGv; E: MGm). C, p = 0.3441 (unpaired t test, 1d SE: 5 slices; 1d NE: 5 slices); D, p = 0.7193 (unpaired t test, 1d SE: 5 slices; 1d NE: 5 slices); E, p = 0.3843 (unpaired t test, 1d SE: 5 slices; 1d NE: 5 slices). F–G, Representative images of coronal brain slices showing calbindin immunostaining (magenta) and ChR2 fluorescence (green) from 10d SE (F) and 10d NE (G). H–J, Average ChR2 immunofluorescence intensity in the different MGB subdivisions at 10d SE and 10d NE (H, MGd; I, MGv; J: MGm). H, p = 0.5064 (unpaired t test, 10d SE: 5 slices; 10d NE: 5 slices); I, p = 0.7387 (unpaired t test, 10d SE: 5 slices; 10d NE: 5 slices); J, p = 0.1008 (unpaired t test, 10d SE: 5 slices; 10d NE: 5 slices). Detailed statistical values are listed in Table 1.
Enhanced activity-dependent facilitation in CT→MGd, but not CT→MGv synapses after NIHL
In the somatosensory cortex, under normal (sparse, or baseline or low) frequency of CT firing, CT→ventral posterior medial nucleus (VPm, core-type thalamus) synaptic activity leads to a transient increase in thalamic neuronal firing rates, which is followed by longer period of robust suppression (Crandall et al., 2015). However, when CTs fire at moderately higher frequencies, such as 5–10 Hz, thalamic neurons show enhanced spiking activity. This transformation is due to facilitation of excitation in CT→VPm synapses and a depression of thalamic reticular nucleus (TRN)→VPm inhibition (Crandall et al., 2015). This corticothalamic switch is important for sensory processing, as it acts as an activity-dependent gating mechanism of sensory input inflow to cortex. Thus, we investigated whether there is an activity-dependent facilitation in AC CT→MGd/v synapses following different frequencies of CT stimulation and whether this facilitation is affected after NIHL. To test this, we delivered a train of 10 brief blue light pulses at frequencies of 1, 5, and 10 Hz, respectively. The peak amplitude of the evoked EPSCs was normalized to the 1st peak in the train for comparison. Consistent with previous studies in somatosensory cortex (Crandall et al., 2015), we observed facilitation of both CT→MGd and CT→MGv synapses at 5 and 10 Hz, but not at 1 Hz (Figs. 6 and 7, SE). Moreover, we did not observe any difference in short-term dynamics between CT→MGd and CT→MGv synapses in SE. Importantly, although we did not find any changes in short-term dynamics in L6 CT→MGv synapses either 1d or 10d after NE versus SE (Fig. 7), we found increased facilitation in CT→MGd synapses in NE versus SE both in 1d (Fig. 6C) and 10d after NE (Fig. 6E). To explore the interaction of synaptic and intrinsic properties in shaping STP and the physiological significance of the voltage-clamp findings in CT→MGd synapses, we employed current-clamp mode experiments. Consistent with our recordings in voltage-clamp mode, we found increased facilitation in CT→MGd synapses in NE versus SE at both 1 and 10d after exposure (Fig. 8), further supporting the physiological significance of our findings. Together, these results support that peripheral hearing loss induces synapse-specific changes on STP in CT→MGB synapses, with enhanced STP in CT→MGd but not CT→MGv synapses. Importantly, these results suggest that after hearing loss, while the corticothalamic auditory CT input to the core-type (lemniscal) MGv pathway remains unchanged, there is an activity-dependent preferential enhancement of CT input to the matrix-type (nonlemniscal) MGd, supporting a compensatory mechanism that likely facilitates auditory perception after hearing loss via activation of matrix-type thalamic neurons and higher-level cortical activity (Discussion).
Enhanced CT→MGd activity-dependent facilitation after NIHL. A–C, Average peak amplitudes of EPSCs in the train, normalized to the peak amplitude of the first EPSC in CT→MGd synapses, in response to a train of 10 light pulses at 1 Hz (A), 5 Hz (B), and 10 Hz (C), in 1d SE (black) and 1d NE (red). A, p = 0.6835 (effect of exposure, 1d SE: 9 cells; 1d NE: 9 cells); B, p = 0.2212 (effect of exposure, 1d SE: 8 cells; 1d NE: 10 cells); C, p = 0.0224 (effect of exposure, 1d SE: 8 cells; 1d NE: 6 cells). D–F, Average peak amplitudes of EPSCs in the train, normalized to the peak amplitude of the first EPSC in CT→MGd synapses, in response to a train of 10 light pulses at 1 Hz (A), 5 Hz (B), and 10 Hz (C), in 10d SE (black) and 10d NE (red). D, p = 0.3406 (effect of exposure, 10d SE: 8 cells; 10d NE: 7 cells); E, p = 0.0239 (effect of exposure, 10d SE: 7 cells; 10d NE: 6 cells); F, p = 0.5526 (effect of exposure, 10d SE: 9 cells; 10d NE: 8 cells). Detailed statistics values are listed in Table 2.
No changes in a CT→MGv activity-dependent facilitation after NIHL. A–C, Average peak amplitudes of EPSCs in the train, normalized to the peak amplitude of the first EPSC in CT→MGv synapses, in response to a train of 10 light pulses at 1 Hz (A), 5 Hz (B), and 10 Hz (C), in 1d SE (black) and 1d NE (red). A, p = 0.9964 (effect of exposure, 1d SE: 4 cells; 1d NE: 8 cells); B, p = 0.9781 (effect of exposure, 1d SE: 6 cells; 1d NE: 10 cells); C, p = 0.5555 (effect of exposure, 1d SE: 7 cells; 1d NE: 11 cells). D–F, Average peak amplitudes of EPSCs in the train, normalized to the peak amplitude of the first EPSC in CT→MGv synapses, in response to a train of 10 light pulses at 1 Hz (A), 5 Hz (B), and 10 Hz (C), in 10d SE (black) and 10d NE (red). D, p = 0.4942 (effect of exposure, 10d SE: 8 cells; 10d NE: 11 cells); E, p = 0.7611 (effect of exposure, 10d SE: 6 cells; 10d NE: 11 cells); F, p = 0.8937 (effect of exposure, 10d SE: 8 cells; 10d NE: 11 cells). Detailed statistical values are listed in Table 2.
Enhanced CT→MGd activity-dependent facilitation after NIHL in current-clamp mode. A–C, Average trough-subtracted peak EPSP amplitudes, normalized to the peak amplitude of the first EPSP in CT→MGd synapses, in response to a train of 10 light pulses at 1 Hz (A), 5 Hz (B), and 10 Hz (C), in 1d SE (black) and 1d NE (red). Peak-trough value was obtained by subtracting the peak amplitude of the EPSP from the through amplitude of the preceding EPSP. The first value is not a peak-trough value; it is the normalized value of the first EPSP peak amplitude. A, p = 0.6477 (effect of exposure, 1d SE: 6 cells; 1d NE: 5 cells); B, p = 0.0102 (effect of exposure, 1d SE: 6 cells; 1d NE: 5 cells); C, p = 0.0076 (effect of exposure, 1d SE: 6 cells; 1d NE: 5 cells). D–F, Average trough-subtracted peak EPSP amplitudes, normalized to the peak amplitude of the first EPSP in CT→MGd synapses, in response to a train of 10 light pulses at 1 Hz (A), 5 Hz (B), and 10 Hz (C), in 10d SE (black) and 10d NE (red). D, p = 0.7744 (effect of exposure, 1d SE: 5 cells; 1d NE: 6 cells); E, p = 0.0270 (effect of exposure, 1d SE: 5 cells; 1d NE: 7 cells); F, p = 0.0130 (effect of exposure, 1d SE: 5 cells; 1d NE: 7 cells).
Discussion
Here, we used Ntsr1-Cre mice to study the synaptic properties of L6 CT neurons. Previous results indicate that Ntsr1+ CT neurons reside in upper L6 (L6a) and lower L6 (L6b; Olsen et al., 2012; Chevee et al., 2018), with the L6a CTs projecting mainly to core-type and L6b CTs projecting mainly to matrix-type thalamus (Bourassa and Deschenes, 1995; Llano and Sherman, 2008; Chevee et al., 2018; Hoerder-Suabedissen et al., 2018; Zolnik et al., 2020; Antunes and Malmierca, 2021). Although our studies did not determine the exact sublayer origin of the CT neurons, our results showing that CTs project to MGv and MGd are consistent with previous studies from somatosensory and motor cortices, which show that Ntsr1+ CTs project to core- and matrix-type thalamus (Yamawaki and Shepherd, 2015; Chevee et al., 2018; Frandolig et al., 2019; Guo et al., 2020; Antunes and Malmierca, 2021; Shepherd and Yamawaki, 2021).
In the somatosensory cortex, CTs project more strongly to the core- versus matrix-type thalamocortical neurons (Guo et al., 2020), but in motor cortex CTs project stronger to matrix- versus the core-type thalamocortical neurons (Harris and Shepherd, 2015; Shepherd and Yamawaki, 2021). However, the auditory cortical CT innervation and synaptic strength in the different subdivisions of the MGB have not been fully determined. Here, we found larger EPSC amplitude in CT→MGv compared with CT→MGd synapses, which is consistent with the somatosensory cortex findings. One limitation of our approach is that we do not know the precise localization of CTs within the different subfields of the AC. Based on previous anatomical studies in the AC, the CT→MGv input is likely originating from A1 (Llano and Sherman, 2008), which is tonotopically organized and has relatively simple receptive fields (lemniscal pathway). The CT→MGd input is likely arising from the secondary AC (AII) and a dorsoposterior region (DP; Llano and Sherman, 2008), both of which have weaker tonotopy and more complex receptive fields (nonlemniscal pathway).
In the somatosensory system, the effect of CT neuronal feedback on thalamus is activity dependent. Namely, low-frequency (0.1 Hz) CT stimulation leads to a small transient excitation in the somatosensory thalamus (VPm) that is followed by a strong inhibition and reduced spiking (Crandall et al., 2015). This happens because the relatively weak CT excitatory input to VPm is followed by the strong disynaptic inhibitory input via the TRN, which sends inhibition to the VPm. On the contrary, higher-frequency (5–10 Hz) CT stimulation leads to strong excitation and enhanced VPm spiking (Crandall et al., 2015). This activity-dependent switch is due to the differences in the STP at higher frequencies of CT synapses to TRN and the primary sensory thalamic nucleus. Namely, TRN→VPm synapses exhibit short-term depression, while CT→VPm synapses exhibit short-term facilitation, thereby altering the balance of excitatory and inhibitory inputs to the VPm driven by CT neurons. The visual system displays similar dynamics (Jurgens et al., 2012; Whitt et al., 2022). Here, we found that AC also follows similar dynamics, at least in terms of the STP of the CT→MGv and CT→MGd synapses, which show no plasticity at 1 Hz (Figs. 6A,D, 7A,D, 8A,D) but short-term potentiation at 5 and 10 Hz (Figs. 6B,C,E,F, 7B,C,E,F, 8B,C,E,F). Although we did not test either the short-term dynamics of inhibitory inputs (TRN→MGB) or the intrinsic properties of MGB neurons, such as spike threshold, which could also affect MGB activity, our results support that sensory cortices can gate their own sensory input via an activity-dependent mechanism that controls L6 corticothalamic input.
In the sensory cortex, it has been proposed that the CT projection to core-type thalamus participates in reciprocal cortico-thalamo-cortical circuit and modulates sensory processing, while the CT projection to the matrix-type thalamus likely participates in higher-order functions, such as perception, cognition, and attention (Llano and Sherman, 2008; Homma et al., 2017; Hoerder-Suabedissen et al., 2018; Zolnik et al., 2020; Antunes and Malmierca, 2021; Ibrahim et al., 2021). However, the synaptic mechanisms of these two pathways are largely unknown in the auditory system and especially in hearing loss. In this context, the most important finding of our study is that after NIHL we observed an enhanced facilitation at 5 and 10 Hz CT activation in CT→MGd but not CT→MGv synapses. It has been known that the reduction of auditory nerve input to the brain after noise trauma induces robust plasticity in the AC, and other brain regions of the auditory system, that compensates for the reduced sensory input (Qiu et al., 2000; Auerbach et al., 2014; Chambers et al., 2016; Kumar et al., 2023). However, our findings uncover a previously unknown corticothalamic plasticity mechanism after noise trauma that highlights enhanced CT input to MGd after noise trauma. Given that MGd is a matrix-type thalamic nucleus that projects to higher-order cortical areas that can generate transthalamic activation from lower- to higher-order cortical areas (Sherman, 2016), this mechanism might enhance perceptual recovery after noise trauma via the enhanced involvement of the MGd feedback and higher cortical areas that modulate perception and cognition (Sherman, 2016; Shepherd and Yamawaki, 2021).
Alternatively, or additionally, the noise trauma-induced enhancement of the MGd output that terminates in the L1 apical tuft dendrites of L5 pyramidal neurons would also enhance perceptual recovery after hearing loss, according to the back-propagation activated coupling model (Jones, 1998, 2001; Llinas et al., 2005; Larkum, 2013). This model does not depend on transthalamic activation across different areas but, instead, on the integration of information from thalamocortical and corticothalamic pathways at the cellular level, namely, at L5 pyramidal neurons. According to this model, enhancement of the perceptual recovery would be achieved by enhanced integration of matrix-type MGd and higher-cortex input, which conveys internal representation information onto the L1 apical dendrites of L5 pyramidal neurons (Larkum, 2013), with the core-type MGv input, which conveys external sensory information onto the basal dendrites of the same L5 pyramidal neurons (Larkum, 2013). The enhanced MGd feedback after noise trauma would likely enhance perceptual recovery via enhanced contextual modulation of perception after NIHL and thus enhancement of prediction (Jones, 1998, 2001; Super et al., 2001; Llinas et al., 2005; Larkum, 2013). Importantly, recent data support that L5 dendritic activity in the AC mostly correlates with behavioral actions rather than sensory processing (Ford et al., 2024). Thus, the noise trauma-induced enhancement of the MGd output, which terminates in the L1 apical tuft dendrites of L5 pyramidal neurons, could also enhance perceptual recovery via enhanced behavioral rather than sensory processing performance.
One limitation of our study is that our experiments were conducted in 2 mM external Ca2+. While we understand that 1.2 mM external Ca2+ is more closely related to in vivo conditions (Borst, 2010; Forsberg et al., 2019), the main finding of our study showing enhanced synaptic facilitation in CT→MGd synapses after noise trauma is likely underestimated in our recording conditions. An additional limitation of our study is that we have not investigated yet the mechanisms underlying the rate-dependent increase in CT→MGd synapses after noise trauma. One potential mechanism would be that noise trauma lowers the threshold for recruiting a reserve pool of vesicles (Denker and Rizzoli, 2010) during the train stimuli. Moreover, changes in the recruitment of neuromodulatory systems during trains of activity after noise trauma, such as endocannabinoid signaling (Kalappa and Tzounopoulos, 2017; Whitt et al., 2022), might be involved in the observed changes in synaptic dynamics. In future studies, we plan to explore the precise synaptic mechanisms underlying this noise-induced plasticity in the synaptic dynamics of CT→MGd synapses.
When we employed voltage-clamp mode to assess synaptic facilitation in CT→MGd synapses in NE and SE (Fig. 6), the difference between NE and SE is significant at 1 but not 10d post exposure for 10 Hz stimulation, suggesting that synapses might be recovering during this period. On the other hand, the difference is significant for 5 Hz stimulation at 10 but not 1d after NE, suggesting that compensatory changes are growing over this period (Fig. 6). However, when we employed current-clamp mode (Fig. 8), which is closer to the physiological conditions, we found increased synaptic facilitation at 5 and 10 Hz stimulation both at 1 and 10 d after noise trauma (Fig. 8). This result supports that the observed plasticity is stable for at least 10 d after noise trauma.
It is important to note that we cannot exclude the possibility that the plasticity described here might contribute to tinnitus, where internally generated pre-percepts, which are likely stored in our brain for predictive purposes, can be released involuntarily (Henton and Tzounopoulos, 2021). In this context, the enhanced feedback inputs after hearing loss could contribute to tinnitus. Taken together, further understanding of the synaptic mechanisms underlying plasticity in cortico-thalamo-cortical loops after peripheral trauma will not only elucidate cortical plasticity mechanisms after sensory organ damage but also holds the potential to highlight novel targets that may either enhance perceptual recovery after sensory organ damage or mitigate plasticity-related sensory processing disorders, such as tinnitus, hyperacusis, and phantom limb pain.
Footnotes
This work was supported by National Institutes of Health grant awards R01-DC019618 and R01-DC020923 (to T.T.).
The authors declare no competing financial interests.
- Correspondence should be addressed to Thanos Tzounopoulos at thanos{at}pitt.edu or Yanjun Zhao at yanjunz{at}pitt.edu.
