Conditional mGluR5 Knockout in Glutamatergic Pathways Disrupts the Development of Excitatory Synaptic Transmission onto Mouse MNTB Neurons

Abstract

Metabotropic glutamate receptor 5 (mGluR5) plays a pivotal role in neurodevelopment. Here, we investigated the consequences of mGluR5 loss-of-function on the development of glutamatergic transmission onto the medial nucleus of the trapezoid body (MNTB). Using the Cre-loxP system, we generated a conditional knockout (KO) mouse line in which mGluR5 expression was selectively eliminated in vesicular glutamate transporter 2 (VGluT2)-expressing glutamatergic pathways, including the calyx of Held synapse innervating MNTB neurons. Whole-cell patch-clamp recordings from mice of either sex at postnatal days 30–38 were used to compare the excitatory synaptic properties of MNTB neurons between KO mice and wild type controls. Upon afferent stimulation of the trapezoid body, MNTB neurons exhibited two distinct types of evoked EPSCs (eEPSCs): large calyceal all-or-none and smaller non-calyceal responses. In mGluR5 KO mice, there was a significant increase in the proportion of neurons exhibiting non-calyceal eEPSCs. The calyceal all-or-none eEPSCs showed significantly prolonged latency, along with slower kinetics in both eEPSCs and asynchronous EPSCs. Analysis of short-term synaptic plasticity of the non-calyceal eEPSCs revealed an increased paired-pulse ratio in mGluR5 KO mice. Additionally, membrane capacitance was significantly reduced, consistent with a smaller somatic area in mGluR5 KO mice. These results suggest that mGluR5 plays a critical role in shaping the excitatory synaptic properties necessary for fast temporal processing in the MNTB.

NEW & NOTEWORTHY

mGluR5 is known to play critical roles in neurodevelopment, but its specific contribution to auditory circuit formation has remained unknown. Using a conditional mGluR5 knockout mouse model, we show that a major glutamatergic pathway in the auditory brainstem is impaired, particularly in synaptic timing, and is accompanied by a reduced somatic area of the postsynaptic neurons. These findings highlight a pivotal role for mGluR5 in shaping auditory brainstem circuitry.

Graphical Abstract

INTRODUCTION

Glutamatergic transmission plays a key role in the development of sensory systems, in part through neuromodulation mediated by metabotropic glutamate receptor 5 (mGluR5), a member of group I mGluRs, via distinct mechanisms (1–3). Dysregulation of mGluR5 has been implicated in several neurodevelopmental disorders, including Fragile X syndrome (4), making mGluR5 a promising target for drug development aimed at treating such brain disorders (5–9). Although mGluR5 expression has not been reported in the auditory peripheral organ of Corti, this receptor is widely expressed throughout the central auditory system (10, 11). However, the role and underlying mechanisms of mGluR5 regulation in auditory system development remain largely unclear.

Neuromodulation at the calyx of Held–MNTB synapse via G protein-coupled receptor (GPCR) systems has been extensively studied (12). Recent studies have expanded our understanding of modulatory pathways at this synapse, including cholinergic modulation (13, 14). A comprehensive understanding of neuromodulation in the MNTB, however, remains incomplete and warrants further investigation. We recently reported that activation of group I mGluRs enhances spontaneous synaptic transmission in the MNTB (15, 16) and, intriguingly, induces temporally patterned transmitter release in a subpopulation of neurons (17). This patterned release appears to be action potential–driven and correlates with the temporally patterned spike activity generated by mGluRs in presynaptic neurons (17). This finding raises an important question about how such centrally generated, patterned spike activity influences the development of mGluR5-expressing neural circuits.

To date, the developmental auditory phenotypes associated with dysfunctional mGluR5, and their underlying mechanisms remain completely unknown. Previous studies—including our own—focused on investigating mGluR5 modulation in models with intact receptor expression (10, 11, 18). In the current study, we investigated which developmental aspects of excitatory transmission onto the MNTB are disrupted by a targeted deletion of mGluR5. Specifically, we focused on identifying cellular alterations resulting from the loss of mGluR5 in glutamatergic pathways. We suggest that mGluR5 is essential for the development of structural and functional features critical for fast temporal processing. To test our hypothesis, we generated a conditional knockout (KO) mouse line in which mGluR5 was selectively eliminated from glutamatergic pathways using the Cre-loxP system. We then examined excitatory synaptic transmission onto MNTB neurons in mice aged postnatal days 30–38 (P30–38), a developmental stage at which cellular and circuit-level maturation is considered complete (19).

MATERIALS AND METHODS

Ethical Approval

All animal procedures were approved by the Institutional Animal Care and Use Committees at the Northeast Ohio Medical University (NEOMED) and Florida State University (FSU) and were conducted in accordance with the National Institutes of Health policies on animal use. VGluT2-Cre mice (Jackson Laboratory, Stock#: 028863) and mGluR5 floxed mice (Jackson Laboratory, Stock#: 028626) were purchased from the Jackson Laboratory and bred at NEOMED. Genotype was confirmed with standard PCR protocol provided by the Jackson Laboratory. Genotyping was performed by Transnetyx (Cordova, TN). Mice were housed in a vivarium with a normal light-dark cycle (12 h light and 12 h dark).

Generation of mGluR5 KO Mice Using Cre-loxP System

Three versions of VGluT exist in the central nervous system (CNS) (20). In adult mice, VGluT1 and VGluT2 are dominant in subcortical structures (21). In the brainstem sound localization circuit, VGluT2 is expressed in glutamatergic neurons including bushy cells in the anteroventral cochlear nucleus (AVCN) (22) and is localized in their axonal terminals including the calyx of Held situated in the MNTB (23, 24). We established and validated floxed mGluR5 and VGluT2-Cre transgenic mouse lines (Fig. 1A–E). Crossing these lines with a ROSA26 reporter (Jackson Laboratory, Stock#: 007676; B6.129(Cg)-Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)Luo}/J) generated F1 heterozygous offspring, which were bred to obtain mGluR5 KO and littermate controls. The Cre-reporter allele labels Cre-expressing cells with membrane-localized EGFP (green) and non-Cre-expressing cells with red fluorescence. We confirmed that in WT mice, there was an absence of Cre activity and EGFP expression (Fig. 1B). In the KO, EGFP was detected in the incoming axons and the calyx of Held in the MNTB (Fig. 1C). These axons are presumably the glutamatergic fibers originating from the contralateral AVCN. In WT mice, double labeling of VGluT2 and mGluR5 confirmed the presence of mGluR5 in both presynaptic calyx terminals and postsynaptic MNTB neurons (Fig. 1D), consistent with our previous report (16). In KO mice, mGluR5 in the VGluT2-containing presynaptic structures in MNTB was diminished (Fig. 1E), whereas mGluR5 in the glycinergic postsynaptic MNTB neurons was preserved. These results validate the effectiveness and specificity of genetic elimination of mGluR5 using the Cre-loxP system.

Figure 1. Knockout (KO) of mGluR5 on VGluT2-expressing glutamatergic pathways reduces mouse body weight and shifts eEPSC pattern in MNTB neurons.

A. A simplified diagram showing mouse breeding strategy. B-C. Validation of mGluR5 KO in VGluT2-expressing neurons. The reporter gene ROSA26 produced tdTomato (red) in non-Cre-expressing neurons in the WT (B). In the KO mouse, EGFP (green) is seen in glutamatergic axons (C), indicating Cre activity. The cell bodies (*) are void of fluorescence because this reporter gene product is membrane-bound. D-E. Double labeling of mGluR5 (green) and VGluT2 (red) in WT and KO littermates without crossing with ROSA26, showing diminished mGluR5 in the calyx terminals (white arrows) of the KO (D, E). As expected, mGluR5 within the MNTB neurons remained because MNTB neurons are glycinergic and lack VGluT2. F. Timeline of mouse auditory development after birth. G. KO mice exhibited lower body weight (N_WT = 24, N_KO = 19, Mann-Whitney test, p = 0.0039). H, I. Representative eEPSCs with calyceal all-or-none response pattern, characterized by reaching the maximal response amplitude (nA range) once the stimulus intensity is suprathreshold. J, K. Representative eEPSCs with non-calyceal response pattern, characterized by increasing amplitude with increasing stimulus intensity and much smaller maximal amplitude (pA range). L. KO mice exhibited proportionally a significantly higher number of MNTB neurons with non-calyceal eEPSCs (Chi-square test, p = 0.0267). In this and subsequent figures, mean and standard error of the mean (SEM) are reported, unless indicated otherwise. The significance levels are indicated by * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001.

Brain Slice Preparation and in vitro Whole-cell Recordings

Coronal brainstem slices (~250 μm in thickness) were prepared from P30–38 old mice of both sexes, as described in our previous publication (25). Mice were deeply anesthetized with isoflurane. Their body weight was measured followed by rapid decapitation. The brainstem was removed and sliced in ice-cold low-Na⁺ artificial cerebrospinal fluid (ASCF) containing the following (in mM): 250 glycerol, 3 KCl, 1.2 KH₂PO₄, 20 NaHCO₃, 3 HEPES, 0.12 CaCl₂, 5 MgCl₂, and 10 glucose, pH 7.4 (when gassed with 95% O₂ and 5% CO₂) and osmolarity of ~290 mOsm/L. Slices were incubated in an interface chamber at 34–36 °C for ~1 h in normal ACSF containing the following (in mM): 130 NaCl, 3 KCl, 1.2 KH₂PO₄, 20 NaHCO₃, 3 HEPES, 1.2 CaCl₂, 1.3 MgSO₄, and 10 glucose, pH 7.4 and osmolarity of ~290 mOsm/L. For recording, slices were transferred to a 0.5 mL chamber mounted on a Zeiss Axioskop 2 FS Plus microscope with a 40× water-immersion objective and infrared differential interference contrast optics. The chamber was continuously superfused with ACSF (3–5 mL/min) by gravity.

Patch pipettes were drawn on a PP-830 or PC-100 Microelectrode Puller (Narishige) to a 1–2 μm tip diameter using borosilicate glass micropipettes (inner diameter, 0.84 mm; outer diameter, 1.5 mm, World Precision Instruments). The electrodes had resistances between 3 and 6 MΩ when filled with internal solution containing the following (in mM): 130 K-gluconate, 4.5 MgCl₂, 4.4 Tris-phosphocreatine, 9 HEPES, 5 EGTA, 4 Na-ATP, 0.48 Na-GTP, with pH 7.3, adjusted with KOH, and osmolarity ~ 290 mOsm/L. QX-314 (5 mM) was added to the internal solution to block Na_v channels and action currents. The liquid junction potential was 9.7 mV, and data were corrected accordingly. The MNTB can be readily identified, and principal MNTB neurons clustered in the middle and lateral parts of the nucleus were selected physiological recordings for both the WT and KO mice. Voltage-clamp experiments were performed with an AxoPatch 200B amplifier (Molecular Devices). Recordings were performed under near physiological temperatures (34–36 °C) and at a holding potential of −70 mV. Data were low-pass filtered at 5 kHz and digitized with a Data Acquisition Interface ITC-18 (InstruTech) at 50 kHz. Recording protocols were written and run using the acquisition and analysis software AxoGraph X (AxoGraph Scientific).

After establishing whole-cell configuration, EPSCs were isolated pharmacologically with bath application of the antagonists for GABA_A receptors (10 μM gabazine) and glycine receptors (1 μM strychnine). All chemicals were purchased from Sigma-Aldrich except for gabazine which was obtained from Tocris Bioscience (R&D Systems).

Analyses of Electrophysiological Data

The resting membrane potential (RMP) was immediately read once the whole-cell recording mode was established. Cell membrane capacitance (C_m) was read from the whole-cell parameter dial of the 200B amplifier. The input resistance (R_in) was calculated from the steady-state current response to a −5 mV test pulse based on Ohm’s Law. For electrical stimulus-evoked EPSCs (eEPSCs), the peak amplitude was measured from the baseline immediately prior to the stimulus artifact and the latency as the time between the stimulus onset and the eEPSC peak. The kinetics analyses include rise time (10–90% to the peak) and decay time constant (tau) which was obtained by a single exponential fitting from the average eEPSC peak to the steady-state baseline.

Spontaneous EPSCs (sEPSCs) were detected by a template using a function for product of exponentials, f(t)=[1 – exp (−t/rise time)] × exp(−t/decay tau), where t stands for time and tau for time constant. The values of the parameters for the template to detect sEPSCs are amplitude of −30 pA, rise time of 0.3 ms, decay tau of 0.5 ms, with a template baseline of 1 ms and a template length of 3 ms. These parameters were determined based on the averaged trace of visually detected synaptic events. The detection threshold is typically 2.5-fold the noise standard deviation, which detects most of the events with the least number of false positives. These same methods were also used for asynchronous EPSC (aEPSC) detection (within 100 ms time window following the termination of a 100 Hz train stimulation) and analyses.

Statistical analyses were performed, and graphs were made with GraphPad Prism10 (GraphPad Software). Data from age-matched and Cre-negative control littermates were compiled and referred as the wild type (WT). Mean ± SEM (standard error of the mean), the number of cells (n), and the number of animals (N) are reported, unless otherwise indicated. Data were subject to Kolmogorov-Smirnov test to examine the normality of distribution, followed by Chi-square test, un-paired t-test, Mann-Whitney test, and Mixed-effects analysis. Test p < 0.05 is considered statistically significant.

Double Labeling of VGluT2 and mGluR5

After physiological recordings, some brainstem slices were fixed in 4% PFA for overnight, and then rinsed with PBS. The fixed brainstem slices were cryoprotected in 30% sucrose in 0.1M phosphate buffer and sectioned at 80 μm using a freezing sliding microtome. Sections were collected in 0.01 M phosphate-buffered saline (PBS) and incubated overnight at 4°C with the primary antibodies anti-VGluT2 (Synaptic Systems Cat# 135 421, RRID:AB_2619823; 1:1000) and anti-mGluR5 (Abcam Cat# ab76316, RRID:AB_1523944; 1:500), diluted in PBS containing 0.3% Triton X-100 and 5% normal goat serum. After PBS washes, sections were incubated overnight at 4°C with Alexa Fluor-conjugated secondary antibodies (Thermo Fisher Scientific Cat# A-11019, RRID:AB_143162; Cat# A-11008, RRID:AB_143165) at 1:1000 in PBS. Following additional PBS washes, sections were mounted on gelatin-coated slides and cover-slipped with Fluoromount-G (SouthernBiotech, Birmingham, AL). Confocal images of the MNTB were captured using a 20× and 60× objectives on an Olympus FV-1200 confocal microscope (Olympus, Breinigsville, USA).

MNTB Somatic Area Measurement

Two groups were used for MNTB body size measurements. In the first group, to measure MNTB somatic area in WT mice which have the same genetic background as the mice used in the Cre-loxP system, fixed brainstem sections were prepared as described previously (26). Animals were deeply anesthetized with an intraperitoneal injection of ketamine and xylazine and transcardially perfused with 0.9% saline followed by 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer. Brains were removed from the skull, post-fixed in 4% PFA overnight at 4°C, and transferred to 30% sucrose for cryoprotection. Brains were sectioned coronally at 40 μm and collected in PBS. Sections were stained with NeuroTrace 640/660 (Thermo Fisher Scientific Cat# N21483, RRID:AB_2572212; 1:1000), a fluorescent Nissl stain, for cell body labeling.

In the second group, to compare MNTB somatic area between WT and KO, fixed slices containing physiologically recorded MNTB neurons were assigned a random numerical identifier to blind the experimenter to genotype and subject identity until statistical analyses were performed. Slices were cryoprotected, sectioned, immunostained, and coverslipped as described above. The primary antibodies included anti-VGluT2 (Synaptic Systems Cat# 135 421, RRID:AB_2619823; 1:1000) and anti-MAP2 (Millipore Cat# MAB3418, RRID:AB_94856; 1:1000). During secondary antibody incubation, NeuroTrace 640/660 was included to label all cell bodies.

MNTB principal neuron size was assessed by measuring the cross-sectional somatic area of individual cells. For each animal, both MNTBs from a single brainstem section were selected. Confocal images were captured using a 20× objective on the Olympus FV-1200 confocal microscope and processed in Fiji software (National Institutes of Health; ImageJ2). Each MNTB was subdivided into medial and lateral halves along the tonotopic axis. Neuronal somata were identified using NeuroTrace staining, with MAP2 and VGluT2 staining confirming neuronal identity. Within each subregion, all neurons displaying a clearly defined border and distinct nucleus in the NeuroTrace channel were included. Somata were manually outlined, and cross-sectional somatic areas were measured.

Statistical analyses were conducted using Prism software (GraphPad, La Jolla, CA). Significance was determined by p < 0.05. The effects of tonotopic location and mGluR5 genotype on MNTB somatic area were assessed using two-way ANOVA, treating either neurons or animals as independent data points. Data are presented as mean ± SD (standard deviation). For illustrative purposes, image brightness, contrast, and gamma were adjusted using Adobe Photoshop (Adobe Systems, Mountain View, CA).

RESULTS

Lack of mGluR5 Shifts the eEPSCs towards a More Non-calyceal Response Pattern

To study the role of mGluR5 in the maturation of the excitatory synaptic transmission in MNTB, we compared whole-cell patch clamp recordings from MNTB neurons of WT mice and mGluR5 KO mice that lacked mGluR5 in VGluT2-expressing glutamatergic cells. The experiments were conducted in mice at an age between P30 and P38 (Fig. 1F). Consistent with a previous study of global mGluR5 KO (27), the body weight of KO mice was significantly smaller than that of WT (Fig. 1G, WT, 16.3 ± 1.1 g, N = 24; KO, 12.5 ± 1.2 g, N = 19, Mann Whitney test, p = 0.0039). This phenotype provides additional confirmation of successful genetic deletion of mGluR5.

In response to electrical stimulation of the afferent fibers in the trapezoid body, MNTB neurons exhibited two distinct responses: all-or-none large (nA range) eEPSCs and smaller (pA range) eEPSCs. Following a previous study (28), these responses are termed calyceal all-or-none and non-calyceal eEPSCs. For neurons with calyceal all-or-none responses, once the stimulation intensity reached the threshold, the amplitude of the eEPSCs reached its maximum with minimal fluctuation (Fig. 1H, I). In contrast, for non-calyceal response pattern, the amplitudes of the eEPSCs were much smaller and variable with increasing stimulus intensity (Fig. 1J, K). Of the 31 MNTB neurons recorded from WT mice, 24 (77%) neurons exhibited calyceal all-or-none responses, and 7 (23%) neurons exhibited non-calyceal responses. Of the 21 MNTB neurons recorded from KO mice, 10 (48%) neurons exhibited calyceal all-or-none responses, and 11 (52%) neurons exhibited non-calyceal responses. Proportionally, KO mice had a significantly lower number of calyceal all-or-none responses (Fig. 1L, Chi-square test, p = 0.0267).

Lack of mGluR5 Prolongs the Latency of Calyceal All-or-none eEPSCs

The calyx-MNTB synapse transmits synaptic information with high precision (in μs range) and high reliability. Disruption of the timing of the neurotransmission at this synapse leads to compromised hearing behavior due to disrupted sound localization (29). To examine the roles of mGluR5 in the maturation of neuronal properties of the calyx-MNTB synapse, we analyzed the amplitude and latency of the eEPSCs in response to midline stimulation of the trapezoid body fibers. For neurons with calyceal all-or-none eEPSC responses, the averaged eEPSC amplitude was similar between KO and WT mice (Fig. 2A, B, Mixed-effects analysis, p = 0.8608). However, the latency of the calyceal eEPSCs in KO mice was significantly longer than in WT mice (Fig. 2C, Mixed-effects analysis, p < 0.0001), with unchanged latency jitter (the standard deviation of the latencies of the eEPSCs evoked repeatedly for 12 times at an intensity of 0.2 mA above threshold) (Fig. 2D, WT, n = 20, N = 11, 0.060 ± 0.015 ms; KO, n = 10, N = 4, 0.073 ± 0.032 ms, Mann Whitney test, p = 0.9483). No significant differences were detected in the current threshold (I_th, the minimal stimulus intensity that caused detectable eEPSC) (Fig. 2E, WT, n = 24, 0.21 ± 0.05 mA; KO, n = 10, 0.32 ± 0.07 mA, Mann Whitney test, p = 0.1109), and the maximal eEPSC amplitude (Fig. 2F, WT, n = 24, 5.899 ± 0.517 nA; KO, n = 10, 5.771 ± 1.41 nA, unpaired t test, p = 0.9156). For the kinetics of the eEPSCs, we analyzed the 10–90% rise time and decay tau of the eEPSC evoked at 0.2 mA above the threshold current. The rise time of eEPSCs in KO mice was significantly prolonged (Fig. 2G, WT, n = 24, 0.154 ± 0.009 ms; KO, n = 10, 0.199 ± 0.014 ms, Mann-Whitney test, p = 0.0115), while no changes were observed for decay tau (Fig. 2H).

Figure 2. eEPSCs from KO mice have prolonged response latency.

A. Representative eEPSCs from neurons with calyceal all-or-none responses in WT and KO mice. B, C. The latency of eEPSCs in KO mice was significantly longer than in WT mice (n_WT = 24, N = 11, n_KO = 10, N = 4, Mixed-effects analysis, p < 0.0001), whereas the averaged amplitude was similar (Mixed-effects analysis, p = 0.8608). Stim int re threshold: stimulus intensity relative to the threshold. D. The jitter of the latency was similar between WT and KO. E-H. The rise time of eEPSCs in KO mice was significantly longer than in WT mice (n_WT = 24, n_KO = 10, Mann-Whitney test, p = 0.0115), with similar current threshold, maximal amplitude, and decay tau. I. Representative eEPSCs from neurons with non-calyceal responses in WT and KO mice. J-P. The eEPSCs from neurons with non-calyceal responses were similar between KO and WT mice in amplitude, latency, latency jitter, current threshold, maximal amplitude, rise time, and decay tau (n_WT = 7, N = 5, n_KO = 8, N = 5).

For neurons with non-calyceal eEPSCs, there were no differences between the KO and WT in averaged eEPSC amplitude (Fig. 2I, J, Mixed-effects model (REML), p = 0.4222), eEPSC latency (Fig. 2K, Mixed-effects model (REML), p = 0.7904), the latency jitter (Fig. 2L, WT, n = 7, N = 5, 0.321 ± 0.141 ms; KO, n = 8, N = 5, 0.741 ± 0.187 ms, unpaired t test, p = 0.1032), the I_th (Fig. 2M, WT, n = 6, 0.267 ± 0.049 mA; KO, n =9, 0.367 ± 0.067 mA, Mann Whitney test, p = 0.3327), the maximal amplitude (Fig. 2N, WT, n = 7, 0.437 ± 0.092 nA; KO, n =8, 0.422 ± 0.069 nA, unpaired t test, p = 0.8937), the rise time (Fig. 2O, WT, n = 5, 0.161 ± 0.010 ms; KO, n = 11, 0.130 ± 0.009 ms, unpaired t test, p = 0.0540), and the decay tau (Fig. 2P, WT, n = 5, 0.244 ± 0.047 ms; KO, n = 11, 0.228 ± 0.028 ms, unpaired t test, p = 0.7698). Moreover, the amplitude of eEPSCs of neurons with non-calyceal responses was much smaller compared to calyceal all-or-none responses (Fig. 2A and I). In WT mice, the average maximal amplitude of eEPSCs in neurons with calyceal all-or-none responses was 5.899 nA, whereas it was only 0.437 nA in neurons with non-calyceal responses, and this difference was also observed in KO mice (Fig. 2F and 2N).

Lack of mGluR5 Causes Subtle Changes in Short-term Plasticity

To examine whether mGluR5 affects short-term plasticity of the excitatory synaptic transmission onto the MNTB, we compared eEPSC properties using a paired-pulse paradigm (inter-pulse interval of 10 ms) and a train stimulation (100 Hz, 20 pulses). The variation in the latency and amplitude was minimal in the calyceal all-or-none eEPSCs, but more pronounced in the non-calyceal eEPSCs (Fig. 3A, E). One reason for the large variations in the non-calyceal eEPSC latencies is that double peaks of eEPSCs were evoked in some cells. To keep consistency, we analyzed the largest eEPSCs for their latencies. When we compared the paired-pulse ratio (PPR), calculated by dividing the amplitude of the second eEPSC by that of the first eEPSC, the PPR for neurons with calyceal all-or-none response was similar between WT and KO mice (Fig. 3B, WT, n = 24, N = 11, 1.016 ± 0.020; KO, n = 10, N = 4, 1.071 ± 0.061). However, the PPR for non-calyceal responses in KO mice was significantly larger than in WT mice (Fig. 3F, WT, n = 6, N = 5, 1.478 ± 0.1617; KO, n = 11, N = 5, 3.253 ± 0.558, Mann-Whitney test, p = 0.0477), indicative of synaptic facilitation. Taking advantage of the paired-pulse responses, we analyzed the time period between the two peaks of the eEPSCs. Although the interval between the two eEPSC peaks was similar between the WT and KO regardless the eEPSC response type (Fig. 3C, G), the jitter of the eEPSCs was significantly larger in the KO in animals with non-calyceal responses (Fig. 3H, WT, n = 6, 0.497 ± 0.157 ms; KO, n = 11, 1.378 ± 0.278 ms, unpaired t test, p = 0.0434), suggesting more variable timing of the excitatory input after mGluR5 elimination.

Figure 3. Paired-pulse ratio (PPR) in neurons with non-calyceal eEPSCs increases in KO mice.

A. eEPSCs of calyceal all-or-none response type in WT and KO mice, recorded with paired-pulse stimulation (inter-pulse interval of 10 ms). The averaged traces (darker) are superimposed with individual traces (lighter). B-D. The PPR, peak latency difference, and jitter of the peak latency of neurons with calyceal all-or-none responses were similar between WT and KO mice (n_WT = 23, N = 11, n_KO = 10, N = 4). E. eEPSCs of non-calyceal response type in WT and KO mice. F-H. The PPR of neurons with non-calyceal responses in KO mice was significantly higher compared to that in WT mice (n_WT = 6, N = 5, n_KO = 11, N = 5, Mann-Whitney test, p = 0.0477), and the jitter for the peak latency difference was larger (unpaired t test, p = 0.0434).

In response to the 100 Hz train stimulation, calyceal all-or-none responses exhibited strong synaptic depression in both WT and KO mice (Fig. 4A); as the stimulus number increased, the amplitude of eEPSCs decreased. After plotting the normalized amplitude and the cumulated amplitude of eEPSCs, we extracted the readily releasable pool (RRP, defined as the value at the Y-axis intercept point from a linear regression to the last 10 data points of the cumulative curve) and the release probability (P_r, calculated by dividing the amplitude of eEPSC1 by RRP), using the SMN (Schneggenburger-Meyer-Neher) method (30). Although this analysis method may underestimate the PPR (31), the systematic error would be cancelled out when comparing between the two genotypes. We found that for neurons with calyceal all-or-none responses the eEPSC1 amplitude, PPR, and P_r were similar between the two genotypes (Fig. 4F–H). For cells with non-calyceal eEPSCs, synaptic facilitation was seen in both WT and KO animals (Fig. 4I–K). Analyses of PPR and P_r were not pursued due to the lack of appropriate approaches (31).

Figure 4. Short-term plasticity of eEPSC in response to 100 Hz stimulation is similar between KO and WT mice.

A. Representative eEPSCs and the cumulative amplitude of two sample neurons with calyceal all-or-none responses evoked with a train stimulation (100 Hz, 20 pulses). In the cumulative eEPSC amplitude plots, the green dashed lines indicate a linear regression to the last 10 points of the curves. The Y-axis intercept value is defined as the readily releasable pool (RRP). B, C. Normalized (to the first eEPSC) and cumulative eEPSC amplitude for neurons with calyceal all-or-none responses in WT mice. D, E. Same plots for KO mice. F-H. No significant differences were detected in the amplitude of the first eEPSC (eEPSC1), readily releasable pool (RRP), and release probability (P_r) for neurons with calyceal all-or-none responses between WT and KO. I-K. The cells with non-calyceal eEPSCs displayed a synaptic facilitation in response to the train stimulation. No measurement of RRP and P_r was pursued due to the lack of appropriate methods for this type of responses.

Lack of mGluR5 Has Minimal Effects on Asynchronous and Spontaneous Glutamate Release

Following the termination of the 100 Hz train stimulation, asynchronous glutamate release events (aEPSC) were observed in MNTB neurons (Fig. 5A–D). In neurons with calyceal all-or-none responses, no significant changes were detected in the frequency, amplitude, and rise time of aEPSCs (Fig. 5E–G), whereas the decay tau was significantly increased in KO mice (Fig. 5H, WT, n = 22, N = 11, 0.162 ± 0.014 ms; KO, n = 10, N = 4, 0.244 ± 0.046 ms, unpaired t-test, p = 0.0337). In neurons with non-calyceal responses, no significant differences were detected in these parameters (Fig. 5I–L).

Figure 5. Asynchronous glutamate release in KO mice has slower decay.

A. Representative asynchronous EPSCs (aEPSCs) from neurons with calyceal all-or-none responses. The aEPSCs were detected within a 100-ms window following the termination of a train stimulation (100 Hz, 20 pules), which was repeated 10 times. B. Averaged and normalized (to the peak) aEPSCs for the two sample neurons. C, D. Sample traces and averaged aEPSCs for neurons with non-calyceal responses. E-L. No significant differences were detected between WT and KO among the parameters (frequency, amplitude, rise time, and decay tau) of aEPSCs, except that the decay tau for aEPSCs in neurons with calyceal all-or-none responses was significantly slower in KO mice (unpaired t-test, p = 0.0337).

Representative individual sEPSCs for both calyceal all-or-none and non-calyceal response cells showed no obvious differences, with the averaged traces superimposed closely between WT and KO (Fig. 6A–D). There were no differences in the frequency, amplitude, rise time, or decay tau between the two genotypes (Fig. 6E–L). When comparing sEPSCs between neurons with calyceal all-or-none and neurons with non-calyceal responses, we found that neurons with calyceal all-or-none responses exhibited higher frequency (WT, calyceal all-or-none, n=24, N = 11, 7.225 ± 4.014 Hz; non-calyceal, n=7, N = 5, 1.214 ± 1.278 Hz, Mann Whitney test, p < 0.0001; KO, calyceal all-or-none, n=10, N =4, 7.879 ± 7.106 Hz; non-calyceal, n =10, N = 5, 1.588 ± 3.098 Hz, Mann Whitney test, p =0.0005), with similar amplitude.

Figure 6. Lack of mGluR5 does not affect spontaneous EPSC (sEPSC).

A. Representative sEPSCs of calyceal all-or-none type neurons. B. Averaged and normalized (to the peak) sEPSCs for neurons with calyceal all-or-none responses. C, D. Sample traces and averaged sEPSCs for neurons with non-calyceal responses. E-L. No significant differences were detected between WT and KO among the parameters (frequency, amplitude, rise time, and decay tau) of sEPSCs.

Lack of mGluR5 Hyperpolarizes the RMP and Reduces Somatic Area of MNTB Neurons

To investigate how the lack of mGluR5 affects the passive properties of MNTB neurons, we compared the resting membrane potential (RMP), input resistance (R_in), and membrane capacitance (C_m) between WT and KO mice. For neurons with calyceal all-or-none responses, the RMP in WT mice was not significantly different from that in KO mice (Fig. 7A-1, WT, n = 24, N = 11, −62.7 ± 1.5 mV; KO, n = 10, N = 4, −66.8 ± 1.9 mV, unpaired t-test, p = 0.0863). No significant difference in R_in was detected between WT and KO mice (Fig. 7B-1, WT, n = 24, 140.8 ± 16.3 MΩ; KO, n = 19, 131.8 ± 12.7 MΩ, Mann Whitney test, p = 0.8091). However, membrane capacitance in KO mice was significantly reduced compared to WT mice in neurons with calyceal all-or-none responses (Fig. 7C-1, WT, n = 24, 13.2 ± 0.8 pF; KO, n = 9, 7.4 ± 0.7 pF, unpaired t-test, p = 0.0004). In neurons with non-calyceal responses, the RMP in KO mice was significantly more hyperpolarized than in WT mice (Fig. 7A-2, WT, n = 7, N = 5, −57.3 ± 1.5 mV; KO, n = 11, N = 5, −64.3 ± 1.9 mV, unpaired t-test, p = 0.0073). The R_in in neurons with non-calyceal responses remained unchanged (Fig. 7B-2, WT, n = 7, 135.9 ± 19.2 MΩ; KO, n = 11, 134.9 ± 17.8 MΩ, Mann Whitney test, p = 0.6426). The C_m in neurons with non-calyceal responses was similar between the genotypes (Fig. 7C-2, WT, n = 7, 11.1 ± 1.3 pF; KO, n = 11, 8.3 ± 1.0 pF, unpaired t test, p = 0.1117).

Figure 7. Lack of mGluR5 hyperpolarizes neurons and reduces membrane capacitance (C_m) and somatic area.

A. In the KO, the resting membrane potential (RMP) did not change significantly in neurons with calyceal all-or-none responses (A-1). A significant hyperpolarization in neurons with non-calyceal responses was detected (A-2, n_WT = 7, n_KO = 11, unpaired t-test, p = 0.0073). As a total population, the significant hyperpolarization in the KO remained (A-3, n_WT = 31, n_KO = 21, unpaired t-test, p = 0.0429). B. The input resistance (R_in) remained unchanged regardless of the response type and whether combined or not. C. The membrane capacitance (C_m) of neurons with calyceal all-or-none responses in the KO was significantly smaller (C-1, n_WT = 24, n_KO = 9, unpaired t-test, p = 0.0004). A trend for smaller C_m existed for neurons with non-calyceal responses (C-2), and a significant decrease was detected when combining the two response types (C-3, Mann-Whitney test, p < 0.0001). D. Somatic areas of MNTB neurons in WT mice display a significant difference between the lateral and medial parts. The data points represent the mean somatic area from each MNTB region of each animal (N = 7 mice). E-F. Consistent with the change in C_m, neuronal somatic area in KO mice was significantly smaller. For panel E, the data points represent individual neurons pooled from all cases within each genotype. Sample sizes are noted at the base of each bar. For panel F, the data points represent the mean somatic area from each MNTB region of each animal. The regional difference in somatic area between the medial and lateral part of MNTB was preserved (E and F). Error bars: standard deviation.

Because a positive correlation exists between C_m and membrane area (32), we analyzed somatic area of MNTB neurons. The tonotopic distribution of neuronal properties—gradients of specific parameters along the frequency-coding axis—is a defining feature of auditory neurons (e.g., 33, 34). A previous study reported a gradient in the somatic area of MNTB neurons in the CBA/CaJ background, with neurons located more lateral being progressively larger (35). We confirmed the presence of a similar variation in the C57BL/6 background in which the mGluR5 floxed mice used in this study were raised (Fig. 7D) (WT, P28 mice, N = 7; Wilcoxon match-pairs signed rank test, p = 0.0156). Consistent with reduced C_m in MNTB neurons of KO mice, the somatic area of MNTB neurons in KO mice was significantly reduced (6 WT and 5 KO mice; Fig. 7E–F). For individual cell-based analyses, two-way ANOVA revealed a significant effect of tonotopic location (F _{(1, 224)} = 84.66, p <0.0001) and mGluR5 genotype (F _(1,244) = 120.5, p < 0.0001) on somatic area (Fig. 7E). For individual animal-based analyses, two-way ANOVA confirmed the significant effect of tonotopic location (F _{(1, 18)} = 16.50, p = 0.0007) and mGluR5 genotype (F _(1,18) = 26.85, p < 0.0001) on somatic area (Fig. 7F). Tukey’s multiple comparisons tests reported smaller somatic area in the lateral and medial MNTB of KO mice as compared to the same subregion of WT mice. The regional variation in cross-sectional somatic area was preserved in the KO, suggesting a scaled yet universal effect of mGluR5 on the postsynaptic somatic area. Due to different preparation procedures, data obtained in slices sectioned after transcardial perfusion (Fig. 7D) were not compared directly with those obtained in slices fixed and re-sectioned after physiological recordings (Fig. 7E, F).

DISCUSSION

We report that targeted deletion of mGluR5 in VGluT2-expressing glutamatergic pathways disrupts the excitatory synaptic transmission onto MNTB neurons. In KO animals, we observed a shift in the eEPSC response pattern from the typical calyceal all-or-none profile to a smaller non-calyceal response. Additionally, eEPSCs exhibited prolonged latency and slower kinetics, accompanied by a reduction in the somatic area of MNTB neurons. These alterations highlight the critical role of mGluR5 in the development of excitatory synaptic transmission within the sound localization circuitry.

Disrupted eEPSC Pattern: Calyceal All-or-none vs Non-calyceal Responses

The shift in eEPSC response patterns observed in mGluR5 KO mice may result from altered development of both calyceal and non-calyceal inputs to the MNTB. The AVCN–MNTB synapse is specialized for fast and high-precision temporal coding, achieved through the developmental formation of the unusually large and powerful calyx of Held synapse (36–40). At birth, each MNTB principal neuron receives excitatory input from multiple calyces (41). Synaptic pruning between postnatal days 2 and 4 (P2–P4) leads to largely mono-innervation by P9 (39, 41–43), which persists into maturity and forms a large synaptic terminal enwrapping the postsynaptic soma (44–48). This mature mono-innervating synapse underlies the characteristic calyceal all-or-none eEPSC response. Under pathological conditions (47, 48), delayed pruning results in persistent multi-innervation, where several underdeveloped calyceal terminals innervate a single MNTB neuron. These synapses produce smaller, more variable responses. In addition, MNTB neurons receive non-calyceal excitatory inputs, which generate eEPSCs that are much smaller in amplitude and more variable in timing than those from calyceal inputs (28). Therefore, persistent immature calyceal innervation and/or an increase in non-calyceal input could contribute to the non-calyceal eEPSC response pattern observed in KO mice in the current study, similar to findings in a Fragile X syndrome mouse model (49).

The functions of most auditory targets of the MNTB depend on temporally fast inhibitory input (50). Therefore, the observed shift from a calyceal all-or-none to a non-calyceal eEPSC pattern could impair the cellular properties essential for fast temporal processing. MNTB neurons provide glycinergic inhibition to multiple targets, primarily within the superior olivary complex (SOC), which plays a central role in sound localization. In addition, the MNTB projects to auditory structures upstream of the SOC, including a direct long projection to the auditory thalamus (51). These inhibitory projections may contribute to temporal processing involved in encoding sound spectrum, duration, and speech comprehension (29, 52). Thus, mGluR5 modulation at the level of the MNTB may influence not only synaptic function within the brainstem circuits for sound localization but also extend its impact to multiple subcortical auditory structures.

Compromised EPSC Timing: Prolonged Latency and Slowed Kinetics

Synaptic plasticity under pathological conditions often involves presynaptic ultrastructural remodeling (53). In mGluR5 KO mice, we observed an increase in eEPSC latency by several hundred microseconds at the AVCN–MNTB synapse in neurons with calyceal all-or-none responses. While such a change may seem modest for average brain synapses, it could significantly impair temporal coding in sound localization circuits, where microsecond-level speed is essential. The prolonged latency may result from alterations in multiple synaptic features, including myelination of afferent fibers, synaptic vesicle pool size, vesicle positioning relative to the active zone, and synaptic cleft width (54, 55). Although mGluR5 is expressed in oligodendrocytes during development (56, 57), where it regulates cell proliferation and contributes to myelin formation via downstream kinase signaling (56, 58, 59), our KO model here is not expected to directly affect mGluR5 in oligodendrocytes because VGluT2 is expressed in neuronal but not glial cells. Rather, because auditory experience promotes and maintains axonal myelination (60) and mGluR5 is expressed presynaptically in the calyx of Held (16), it is conceivable that mGluR5 loss-of-function in VGluT2-expressing pathways indirectly leads to reduced myelination of AVCN axons projecting to the MNTB, contributing to the prolonged synaptic transmission between the calyx of Held and MNTB principal neurons. This idea is consistent with a recent study which shows that mGluR5 mediates activity-driven myelin sheath growth (61).

The slower kinetics of the excitatory responses—reflected in the prolonged rise time of eEPSCs and increased decay tau of aEPSCs in neurons with calyceal all-or-none responses—suggest potential alterations in the presynaptic machinery of glutamate release in the KO. By P14, the remarkable shortening of presynaptic action potentials enables MNTB eEPSCs to become extremely fast, allowing them to follow high-frequency inputs of up to 800 Hz (62). We suspect that this shortening of presynaptic action potentials was impaired in the mGluR5 KO animals. Changes in postsynaptic ionotropic glutamate receptors may be excluded because the principal MNTB neurons are glycinergic, absent of VGluT2. Interestingly, the kinetics of sEPSCs did not exhibit a significant slowing, suggesting that the molecular mechanisms governing aEPSCs may differ from those underlying sEPSCs (63).

We observed heterogeneity in the statistical significance of neuronal properties between the two response types in the KO. Significant changes in certain parameters (eEPSC latency and rise time, aEPSC decay tau, and C_m) were detected in cells with calyceal all-or-none responses but not in those with non-calyceal responses. Conversely, other parameters (PPR, eEPSC jitter, and RMP) changed significantly in cells with non-calyceal responses but not in the all-or-none type. Interestingly, for every change observed in one response type, there was at least a trend toward a change in the same direction (increase or decrease) in the other type, despite the lack of statistical significance. These findings suggest that mGluR5 effects on MNTB neuronal properties are consistent in modulatory direction but differ in magnitude across cells. In addition, our recordings were obtained only from animals with a mature auditory system. How neuronal properties differ between WT and KO animals, and whether such heterogeneity exists during development (e.g., around the age of hearing onset), remain unknown and warrant future investigation.

Disrupted Morphology: Reduced Somatic Area

The development of synaptic functions relies on the proper formation of synaptic connectivity, and in many neurodevelopmental disorders, functional deficits are accompanied by structural malformations (64). For instance, loss-of-function of mGluR5 in the cerebral cortex results in less polarized dendritic patterning (65, 66), decreased dendritic length (67), misorientation of dendrites (68), and abnormal spine morphology (2, 69). In the case of the MNTB, how does the postsynaptic neuron shrink in size following mGluR5 knockout from presynaptic neurons? Reduced afferent activity in the calyx has been shown to decrease MNTB somatic area in deaf mice (35). We speculate that mGluR5 in the presynaptic calyx modulates spike activity to the MNTB, and that reduced activity in mGluR5 KO mice impairs the normal growth of MNTB cell bodies during development. Interestingly, the tonotopic pattern of the somatic area was preserved in the mGluR5 KO, suggesting that mGluR5 exerts a scaled, universal effect on postsynaptic MNTB neurons.

Possible Mechanisms and Functional Implications

Homeostatic regulation of neural properties by antagonistic mechanisms is a fundamental aspect of brain function. mGluR5 enhances protein synthesis, counteracting the actions of the Fragile X messenger ribonucleoprotein 1 (FMRP). Although clinical trials targeting mGluR5 for the treatment of FXS have not been successful (70), extensive research has demonstrated that FMRP and mGluR5 interact closely in brain health and disease (71). In animal models, dysfunction in mGluR5 contributes to many FXS symptoms, including abnormal dendritic and spine development in cortical circuits (5). Loss of FMRP leads to elevated mGluR5 activity (72), contributing to FXS phenotypes (73) including sound hypersensitivity, which increases the risk of acoustic seizures and death (74). Therefore, dysfunctional mGluR5 may disrupt neurodevelopment through mechanisms opposite to those resulting from FMRP loss.

DATA AVAILABILITY

The datasets generated for this study are available upon request to the corresponding author.