Transmission of auditory sensory information decreases in rate and temporal precision at the endbulb of Held synapse during age-related hearing loss

This study performed patch-clamp recording in cochlear nucleus slices from young (1–3 mo) and old (25–30 mo) CBA/CaJ mice. For the first time, the study revealed the functional decline of the spiral ganglion cells at their central synapses in the cochlear nucleus during aging, which is expected to contribute to the perceptual deficits of age-related hearing loss.

Abstract

Age-related hearing loss (ARHL) is largely attributed to structural changes and functional declines in the peripheral auditory system, which include synaptopathy at the inner hair cell/spiral ganglion cell (SGC) connection and the loss of SGCs. However, functional changes at the central terminals of SGCs, namely the auditory nerve synapses in the cochlear nucleus, are not yet fully understood during ARHL. With the use of young (1–3 mo) and old (25–30 mo) CBA/CaJ mice, this study evaluated the intrinsic properties of the bushy neurons postsynaptic to the endbulb of Held synapses, and the firing properties of these neurons to direct current injections as well as to synaptic inputs from the auditory nerve. Results showed that bushy neurons in old mice are more excitable and are able to fire spikes at similar rate and timing to direct current injections as those in young mice. In response to synaptic inputs, however, bushy neurons from old mice fired spikes with significantly decreased rate and reduced temporal precision to stimulus trains at 100 and 400 Hz, with the drop in firing probability more profound at 400 Hz. It suggests that transmission of auditory information at the endbulb is declined in both rate and timing during aging, which signifies the loss of sensory inputs to the central auditory system under ARHL. The study proposes that, in addition to damages at the peripheral terminals of SGCs as well as the loss of SGCs, functional decline at the central terminals of surviving SGCs is also an essential component of ARHL.

NEW & NOTEWORTHY

This study performed patch-clamp recording in cochlear nucleus slices from young (1–3 mo) and old (25–30 mo) CBA/CaJ mice. For the first time, the study revealed the functional decline of the spiral ganglion cells at their central synapses in the cochlear nucleus during aging, which is expected to contribute to the perceptual deficits of age-related hearing loss.

age-related hearing loss (ARHL) is a major health impairment in the elderly that affects two-thirds of the U.S. population aged above 70 (Lin et al. 2011a). The primary causes of ARHL have been mostly attributed to functional declines in the peripheral auditory structures (Mao et al. 2013), including decreased endocochlear potential due to changes of the stria vascularis (Schulte and Schmiedt 1992), loss of outer and inner hair cells (Bohne et al. 1990), and loss of spiral ganglion cells (SGCs) (Makary et al. 2011; Sergeyenko et al. 2013; Sha et al. 2008). Particular interest has been focused on SGCs that link the peripheral and central auditory system by transmitting sound information from the cochlea to cochlear nucleus, through connections at SGC peripheral terminals that receive synaptic inputs from inner hair cells, and at SGC central terminals that synapse onto cochlear nucleus neurons (Nayagam et al. 2011). Recent studies showed that SGC peripheral terminals are vulnerable and can be damaged before other peripheral structures during noise trauma (Furman et al. 2013; Kujawa and Liberman 2009; Lin et al. 2011b) or aging (Sergeyenko et al. 2013; Viana et al. 2015). This synaptopathy disrupts the information flow in affected SGCs and is believed to underlie perceptual deficits of both noise-induced and age-related hearing loss, particularly the “hidden hearing loss” that is associated with difficulties in speech recognition and temporal processing (Bharadwaj et al. 2014; Grose and Mamo 2010; Kujawa and Liberman 2015; Plack et al. 2014; Viana et al. 2015). Despite their vulnerability, the majority of SGC peripheral terminals (Sergeyenko et al. 2013; Viana et al. 2015), as well as the SGC themselves, survive through aging (Makary et al. 2011; Sergeyenko et al. 2013; Sha et al. 2008). The function of the surviving SGCs in passing information to the central system also relies on the physiological performance of their central terminals at synapses in the cochlear nucleus. However, it remains unclear to what extent these SGC central terminals maintain their physiological function during aging, and how changes in auditory information transmission across this synaptic interface between SGCs and central neurons in the cochlear nucleus might contribute to the decline in neural processing during ARHL.

One primary central terminal of the SGCs is the endbulb of Held synapse, which is specialized to provide auditory information with fine temporal resolution to the postsynaptic bushy neurons in cochlear nucleus (Manis et al. 2011). Because ARHL is associated with decreased capability in processing temporal fine structure of sound (Anderson et al. 2012; Grose and Mamo 2010; Lorenzi et al. 2006; Plack et al. 2014), it is conceivable that information flow across this endbulb synapse may experience age-related changes that contribute to ARHL. Indeed, our study revealed that synaptic transmission at the endbulb of Held deteriorates during ARHL (Xie and Manis 2016), adversely affecting the effectiveness of signal transmission from the auditory nerve to the postsynaptic bushy neurons.

With the use of young (1–3 mo) and aged (25–30 mo) CBA/CaJ mice, this study investigated physiological changes in how auditory information is passed from SGCs to postsynaptic bushy neurons during ARHL. The results showed that the intrinsic membrane properties are changed in bushy neurons of old mice, which become more excitable than those of young mice. More importantly, bushy neurons in old mice can fire repetitive spikes upon direct current injections with similar rates and timing as those in young mice, suggesting that bushy neurons in old mice are functionally able to encode and pass auditory information from the auditory nerve. However, when the auditory nerve was stimulated to activate synaptic inputs, there was a decline in both firing rate and temporal precision of evoked spikes in bushy neurons of old mice. The decline was more profound at higher firing rates, at which the majority of bushy neurons in old mice appeared unable to fire sufficient spikes to pass sustained information to higher auditory nuclei. The results suggest that, while the ability of bushy neurons to encode auditory information remains relatively normal in old mice, the functional decline at this site is mostly attributable to deteriorated synaptic drive from the endbulbs of Held that innervates them, which was shown by compromised synaptic transmission in old mice (Xie and Manis 2016). The age-associated drop in firing rate and temporal precision of auditory nerve-driven spikes in bushy neurons decreases the signal strength and disturbs the timing of inputs to the central auditory system, and is likely to contribute to the perceptual deficits of ARHL.

MATERIALS AND METHODS

CBA/CaJ mice of either sex were used in two age groups [1–3 mo (young) and 25–30 mo (old)]. Mice were initially purchased from the Jackson Laboratory, bred to establish an onsite mouse colony, and maintained until desired age for experiments. All young mice were housed for their entire lifetime in the animal facility at the University of Toledo Health Science Campus. All old mice were initially housed in the animal facility at the University of North Carolina at Chapel Hill until about 1 yr old and then transferred to the animal facility at the University of Toledo Health Science Campus to continue aging until their use in experiments. The ambient noise level in the housing rooms of both animal facilities was below 70 dB sound pressure level (SPL) in broadband, and below 30 dB SPL at 10 kHz (probed with Larson Davis model 831 sound level meter). All electrophysiological experiments were carried out under the guidelines of the protocols approved by the Institutional Animal Care and Use Committee at the University of Toledo.

Auditory brain stem response.

Before electrophysiological experiments, the auditory brain stem response (ABR) was measured to evaluate the hearing status of each mouse, using the same procedure as previously described (Xie and Manis 2013a, 2016), using a RZ6-A-P1 bioacoustic system with BioSigRZ software (Tucker-Davis Technologies, Alachua, FL). Mice were anesthetized with an intraperitoneal injection of ketamine (100 mg/kg) and xylazine (10 mg/kg) and were placed on a feedback-controlled heating pad to maintain body temperature inside a sound-attenuating box. Clicks (0.1 ms, monophasic with alternating phase) were generated and attenuated at various levels by the RZ6-A-P1 and presented through a free-field magnetic speaker (MF1; Tucker Davis Technologies) located 10 cm away from the ear. ABR signals were collected through needle electrodes that were positioned at the ipsilateral pinna and vertex, with the ground electrode placed at the rump. Signals were amplified by an RA4PA Medusa preamplifier and optically transferred back to RZ6-A-P1 for processing. Clicks were presented 512 times at each sound level at 21 times/s. The averaged ABRs were used for visual determination of hearing thresholds (Xie and Manis 2013a, 2016).

Brain slice preparation.

Following ABR measurement, mice under anesthesia were decapitated, and their brain stems were removed into artificial cerebrospinal fluid (ACSF) for slice preparation (Xie and Manis 2013a, 2013b). The ACSF, which contained (in mM) 122 NaCl, 3 KCl, 1.25 KH₂PO₄, 25 NaHCO₃, 20 glucose, 3 myoinositol, 2 sodium pyruvate, 0.4 ascorbic acid, 2.5 CaCl₂, and 1.5 MgSO₄, was prewarmed to 34°C and gassed with 5% CO₂ and 95% O₂. Parasagittal slices (350 μm) containing the cochlear nucleus were cut using a Vibratome 1000 (Technical Products, St. Louis, MO) and were incubated in ACSF at 34°C for 30 min. For whole cell patch-clamp recording, slices were transferred to a submersion chamber on a fixed-stage Axio Examiner microscope (Carl Zeiss, Oberkochen, Germany), with a continuous flow of gassed ACSF at ∼2–3 ml/min. All recordings were made at 34°C.

Data acquisition.

Current-clamp recordings were obtained from bushy neurons in the high-frequency region of the anteroventral cochlear nucleus using a Multiclamp 700B amplifier, Digidata 1550 acquisition system, and PClamp10 software (Molecular Devices, Sunnyvale, CA). Patch pipettes were pulled from borosilicate glass (KG-33; King Precision Glass, Claremont, CA) on a Sutter P2000 puller (Sutter Instruments, San Francisco, CA). Recordings were low-pass filtered at 6 kHz and digitized at 100 kHz. The electrode solution in the recording pipette contained (in mM) 126 potassium gluconate, 6 KCl, 2 NaCl, 10 HEPES, 0.2 EGTA, 4 MgATP, 0.3 GTP, and 10 Tris-phosphocreatine, with pH adjusted to 7.20. Alexa Fluor 488 was included in the electrode solution to fill the patched neurons for online imaging of cellular morphology. Bushy cells were identified morphologically by having short primary dendrites with heavily branched tufts (Brawer et al. 1974; Cant and Morest 1979; Lauer et al. 2013; Webster and Trune 1982), as well as physiologically by firing only one to a few action potentials to prolonged depolarizing current injections (Wu and Oertel 1984). For current pulse trains (Fig. 3), pulse amplitude was set at such a level that a single test pulse could reliably trigger action potentials in the target bushy neuron, which was about three to four times the threshold current level of the same neuron in Fig. 2, A and B. Electrical stimulation of the auditory nerve (Figs. 4 and 5) was delivered through a 75-μm-diameter concentric stimulating electrode (Frederick Haer, Bowdoin, ME) placed at the auditory nerve root. A single stimulus pulse (100-μs duration) was tested with gradually increasing intensity until reliable action potentials were evoked in the target neuron. The same stimulus intensity was then used for train stimulations of 50 pulses at 100 and 400 Hz. Inhibitory synaptic transmission was blocked by including 2 μM strychnine in the bath ACSF. All membrane potential data were adjusted for a junction potential of −12 mV.

Fig. 3.

Bushy cells from young and old mice can fire spikes to current pulse train injections at 100 and 400 Hz. Red, data from young mice; blue, data from old mice. A: responses to 100-Hz current pulse train injections in example bushy neurons from a young mouse (red) and an old mouse (blue). Notice that bushy neurons from both young and old mice fired spikes reliably to all 50 pulses throughout the train. Black trace, command trace of the current pulse train. B: same as in A except showing responses to 400-Hz current pulse trains. Notice that there were failures during the train in both neurons. C and D: summary of action potential (AP) firing probability per pulse in bushy neurons from young and old mice to 100 (C)- and 400 (D)-Hz pulse trains. E and F: summary of vector strength from evoked spikes to 100 (E)- and 400 (F)-Hz pulse trains. Each symbol represents an individual bushy neuron in C–F.

Fig. 2.

Intrinsic membrane properties of bushy neurons in young and old mice. Red, data from young mice; blue, data from old mice. A: example membrane responses to current step injections in a bushy cell from a young mouse. Top, membrane responses; bottom, steps of current injections; inset, membrane responses to threshold current injection (black, with a triggered action potential) and the next smaller current step injection (gray, failed to trigger any action potential). The resting membrane potential of this bushy neuron was −63.8 mV as marked. B: example membrane responses to current step injections in a bushy cell from an old mouse. Plots are arranged the same as in A. C–F: summary of all bushy cells in resting membrane potential (C), membrane input resistance (D), membrane time constant (E), and threshold current (F). Each symbol represents an individual bushy neuron. NS, not significant; **P < 0.01 and ***P < 0.001.

Fig. 4.

Firing probability and temporal precision of spikes driven by auditory nerve stimulation at 100 Hz. Red, data from young mice; blue, data from old mice. A: responses to auditory nerve stimulation at 100 Hz for an example bushy neuron from a young mouse. Top, single trace response to a stimulus train; bottom left, magnified view of responses to the first two stimuli of the train as marked by the broken lines; bottom right, magnified view of responses to two stimuli later in the train. Panels on bottom include overlap of 12 traces. Notice that evoked spikes overlap with little temporal jitter. Ticks (top) and arrows (bottom) mark the time of the auditory nerve stimulation. B: responses of an example bushy neuron from an old mouse. Plots are arranged the same as in A. Bottom, overlap of 24 traces. Notice that many stimuli during the later stage of the train failed to evoke spikes, and the evoked spikes showed more temporal jitter (bottom right). C: raster plot of spikes evoked by 100-Hz stimulus train from the cell in A. Ticks on top mark stimulus onset. Each dot represents a single spike. Responses to the first 10 stimuli of the train were defined as the onset response and responses to the last 40 stimuli as the sustained response (as marked). D: raster plot of spikes from the bushy cell in B. E: overlapped peristimulus time histogram (PSTH) for spikes from the bushy neuron in A during the onset (left) and sustained phase (right) of the 100-Hz stimulus train. F: overlapped PSTH for spikes from the bushy neuron in B. Notice the broadened distribution of spike timing compared with E. The PSTH bin size is 10 μs in both E and F. G: summary of the action potential firing probability per stimulus during the onset (left) and sustained (right) phase of the 100-Hz trains. H: summary of the vector strength of spikes evoked during the onset (left) and sustained (right) phase of the 100-Hz trains. Each symbol in G and H represents an individual bushy neuron. **P < 0.01 and ***P < 0.001.

Fig. 5.

Firing probability and temporal precision of spikes driven by auditory nerve stimulation at 400 Hz. Red, data from young mice; blue, data from old mice. A: responses to auditory nerve stimulation at 400 Hz for an example bushy neuron from a young mouse. Top, single trace response to a stimulus train; bottom left, magnified view of responses to the first two stimuli of the train as marked by the broken lines; bottom right, magnified view of responses to two stimuli later in the train. Panels on bottom include overlap of 12 traces. B: responses of an example bushy neuron from an old mouse. Plots are arranged the same as in A. Bottom, overlap of 24 traces. Notice that only a few spikes were evoked during the sustained phase of the stimulus train, and the evoked spikes showed prominent temporal jitter (bottom right). C: raster plot of spikes evoked by 400-Hz stimulus train from the bushy cell in A. Ticks on top mark stimulus time. Each dot represents a single spike. D: raster plot of spikes evoked by 400-Hz trains from the bushy cell in B. Notice the decreased spike number in both onset and sustained phase compared with C. E: overlapped PSTH for spikes from the bushy neuron in A during the onset (left) and sustained phase (right) of the 400-Hz stimulus train. F: overlapped PSTH for spikes evoked by 400-Hz trains from the bushy neuron in B. The PSTH bin size is 10 μs in both E and F. G: summary of the action potential firing probability per stimulus during the onset (left) and sustained (right) phase of the 400-Hz trains. Each symbol represents an individual bushy neuron. H: summary of the vector strength in spikes evoked during the onset (left) and sustained (right) phase of the 400-Hz trains. Cells marked with red and blue asterisks fired no or too few spikes; no vector strength was derived for these cells. *P < 0.05 and ***P < 0.001.

Data analysis.

Data analysis was done using Igor pro (version 6.37; WaveMetrics) with custom-written functions. Resting membrane potential was determined as the average membrane potential during a 50-ms window preceding current step injections (Fig. 2, A and B). The membrane time constant was calculated as the average time constant from single exponential fits to small hyperpolarizing traces driven by −25, −50, −75, and −100 pA current injections (fitting range: from onset of the hyperpolarization to the negative peak). Membrane input resistance was calculated from the slope of the current-voltage relationship for the same four hyperpolarizing current injection levels. Spike timing was defined as the peak time of the action potential. Vector strength is a measure of spike synchronization to periodic stimuli (Goldberg and Brown 1969), which ranges from 0 to 1, with 1 representing a perfect synchrony of spike timing to stimulus inputs. To quantify the temporal precision of evoked spikes in bushy neurons from young and old mice, the vector strength r was calculated using the formula: $r={\sqrt{{(\mathrm{\Sigma }\sin {\theta }_i)}^2+{(\mathrm{\Sigma }\cos {\theta }_i)}^2}/n,}^{}$ where θ_i is the phase angle of the spike i relative to the stimulation cycle, and n is the total number of spikes analyzed throughout the stimulus trains.

Statistical analysis.

Statistical analyses were performed using Prism (version 6.0f; GraphPad Software, La Jolla, CA). Measurements from bushy neurons in the young and old mice groups were tested for normality to determine if they were from a Gaussian distribution. The unpaired Student's t-test was then used to determine significant differences between measurements that were normally distributed; otherwise, the Mann-Whitney test was used as stated. Data are presented as median and interquartile range (IQR, the range between first quartile and third quartile) in results and Figs. 1–5.

Fig. 1.

Age-related hearing loss (ARHL) revealed by auditory brain stem response (ABR) in CBA/CaJ mice. A: representative ABR waveforms to clicks of various amplitudes from a young (2-mo-old) and an old (28-mo-old) mouse. Arrows mark the waveform to clicks at threshold intensity. B: ABR waveforms at 85 dB SPL from A. Nos. mark different wave peaks in young mouse. SP, cochlear summating potential. Note that, while SP and wave 1 can be unambiguously identified in the waveform from the old mouse, later waves (2–5) are less clear. C: summary of ABR thresholds to clicks in young and old mice. ***P < 0.001.

RESULTS

Data were collected from 38 young CBA/CaJ mice (1–3 mo; 21 female, 17 male) and 69 old mice (25–30 mo; 39 female, 30 male). Current-clamp recordings were made from 63 bushy cells from the young mice and 84 bushy cells from the old mice.

Old CBA/CaJ mice show ARHL.

To evaluate the hearing status of the CBA/CaJ mice, ABRs were measured to clicks with intensity varied from 20 to 90 dB SPL by 5-dB steps in both young and old mice (Fig. 1). Consistent with our previous observation (Xie and Manis 2016) and other reports (Sergeyenko et al. 2013), CBA/CaJ mice showed significant hearing loss at the age of 25–30 mo. The median ABR threshold to clicks was 30 dB SPL (IQR: 7 dB SPL, n = 38) in young mice and 70 dB SPL (IQR: 16 dB SPL, n = 62) in old mice (Fig. 1C; unpaired t-test: t₉₈ = 19.95, P < 0.0001). Hearing threshold was not determined in seven old mice because no clear ABR waves were observed, even at the highest sound intensity (90 dB SPL) used in the test. As shown in Fig. 1B, distinct cochlear summating potential (SP) as well as five waves were typically observed from ABR recordings in young mice. In old mice, however, only the SP and the first wave could be easily identified, whereas the later waves (waves 2-5) were ambiguous and hard to be distinguished compared with those from the young mice. These later waves are contributed by the neural activity of the central nuclei along the auditory pathway (Buchwald and Huang 1975; Melcher and Kiang 1996). Therefore, the changes in later waves in old mice, including decreased wave amplitude and diminished wave peaks, reflect the modified neural activities in the central auditory system during ARHL.

Intrinsic membrane properties of bushy neurons in young and old mice.

Whole cell current-clamp recordings were obtained from bushy neurons to evaluate their intrinsic membrane properties in young and old mice. After gaining whole cell access to a target neuron, 100-ms current steps from −500 to +500 pA at 25-pA increments were injected to evoke hyperpolarizing and depolarizing membrane potential changes and action potentials (Fig. 2, A and B; only 4 current steps were shown). The resting membrane potential of the bushy neurons from old mice (median: −62.6 mV, IQR: 2.4 mV, n = 84) was significantly more depolarized than that of the bushy neurons from the young mice (median: −63.8 mV, IQR: 2.2 mV, n = 63) (Fig. 2C; Mann-Whitney test: P < 0.0001). The membrane input resistance was significantly higher in bushy cells from old mice (median: 54.5 MΩ, IQR: 21.8 MΩ, n = 84) than those from young mice (median: 47.1 MΩ, IQR: 19.4 MΩ, n = 63) (Fig. 2D; Mann-Whitney test: P = 0.001). There was no significant difference in the membrane time constant of bushy neurons between young (median: 1.94 ms, IQR: 0.78 ms, n = 63) and old (median: 1.87 ms, IQR: 0.90, n = 84) mice (Fig. 2E; Mann-Whitney test: P = 0.728). As shown in Fig. 2, A and B, the threshold current of a bushy cell was determined as the minimum depolarizing current step that triggered an action potential. Bushy neurons from old mice showed significantly smaller threshold current than those from young mice (Fig. 2F; median threshold current in old: 200 pA, IQR: 100 pA, n = 84; young: 325 pA, IQR: 150 pA, n = 63; Mann-Whitney test: P < 0.0001). The changes in the intrinsic membrane properties of bushy neurons in old mice, by having more depolarized resting membrane potentials, higher membrane input resistances, and decreased threshold currents, suggest that these neurons are more excitable and can be triggered to fire action potentials with weaker synaptic drives than bushy neurons in young mice.

Bushy cells in old mice can fire high rates of action potentials to current injections.

The auditory nerve fibers are highly active in transmitting sound information under physiological conditions and can fire action potentials at high rates above 400 Hz (Taberner and Liberman 2005; Wen et al. 2009). The question is whether bushy cells in old mice are capable of firing high rates of action potentials with timing comparable to those in young mice, to encode the information from the auditory nerve for passage to higher auditory nuclei. Trains of 50 suprathreshold current pulses (1-ms duration each) were directly injected in bushy neurons at 100 and 400 Hz to drive repetitive action potentials.

At 100 Hz, bushy neurons from both young and old mice fired action potentials to almost every pulse of the train (Fig. 3A). There was no significant difference in the action potential firing probability per pulse between bushy neurons from young (median: 1.00, IQR: 0, n = 17) and old (median: 1.00, IQR: 0, n = 12) mice (Fig. 3C; Mann-Whitney test: P = 0.163). To quantify the spike timing in bushy neurons evoked by current injection, vector strength was calculated (see materials and methods) from the responses throughout the 100-Hz train. There was no significant difference in the vector strength of bushy cell spikes evoked by 100-Hz pulse trains between young and old mice (Fig. 3E; young median: 0.9998, IQR: 0.0001, n = 17; old median: 0.9998, IQR: 0.0003, n = 12; Mann-Whitney test: P = 0.437). At 400 Hz, bushy neurons from both young and old mice fired action potentials throughout the pulse train with some failures (Fig. 3B). Similar to 100-Hz trains, there was no significant difference in the action potential firing probability of bushy neurons to 400-Hz pulse trains between young (median: 0.81, IQR: 0.28, n = 17) and old (median: 0.89, IQR: 0.36, n = 12) mice (Fig. 3D; Mann-Whitney test: P = 0.308) nor in their vector strength (Fig. 3F; young median: 0.9731, IQR: 0.0347, n = 17; old median: 0.9837, IQR: 0.0214, n = 12; Mann-Whitney test: P = 0.165). The results suggest that, despite the differences in their intrinsic membrane properties, bushy neurons from young and old mice are functionally capable of firing repetitive spikes at high rates up to 400 Hz and with comparable temporal precision.

Action potentials driven by synaptic inputs from the auditory nerve at 100 Hz.

Given that bushy neurons in old mice appear to be functionally normal in firing repetitive spikes to direct current injections, the next question is whether the endbulb of Held synapse in old mice can provide normal synaptic drive to pass auditory information from SGCs to evoke action potentials in bushy neurons. In this experiment, the auditory nerve was electrically shocked (see materials and methods) to trigger synaptic transmission at the endbulbs while recording current-clamp responses in bushy neurons. The firing probability and temporal precision of evoked spikes were assessed to evaluate the function of the endbulb synapses.

For low-frequency inputs, trains of 50-pulse stimulation were delivered at 100 Hz at the auditory nerve. Bushy neurons from young mice reliably fired action potentials to the stimuli throughout the train (Fig. 4, A and C). The spikes were temporally precise with little time jitter both at the beginning and during the later stage of the train (Fig. 4A). Bushy neurons from the old mice fired spikes reliably at the beginning of the 100-Hz train; however, the firing was less robust during the later stage of the train, where stimuli frequently failed to evoke action potentials (Fig. 4, B and D). To better assess spikes evoked at different stages of the trains, responses were grouped into onset responses (first 10 stimuli of the train) and sustained responses (last 40 stimuli of the train) (Fig. 4, C and D). The spikes in bushy neurons from old mice were temporally less precise during both the onset and sustained responses, as shown by the peristimulus time histograms (Fig. 4, E and F). In summary, the action potential firing probability per stimulus during the onset of the 100-Hz trains was similar between bushy neurons from young (median: 1.00, IQR: 0.03, n = 23) and old (median: 1.00, IQR: 0.11, n = 36) (Fig. 4G, left; Mann-Whitney test: P = 0.369) mice. During sustained response, the action potential firing probability was significantly lower in bushy neurons from old mice (median: 0.85, IQR: 0.59, n = 36) than those from young mice (median: 1.00, IQR: 0.05, n = 23) (Fig. 4G, right; Mann-Whitney test: P = 0.0015). For both onset and sustained spikes, the vector strength was significantly smaller in bushy neurons from the old mice than in those from the young mice (Fig. 4H; young onset median: 0.9993, IQR: 0.0015, n = 23; old onset median: 0.9962, IQR: 0.0066, n = 36; Mann-Whitney test: P = 0.0017; young sustained median: 0.9992, IQR: 0.0021, n = 23; old sustained median: 0.9952, IQR: 0.0063, n = 36; Mann-Whitney test: P < 0.0001). These results indicate that, during the onset of the 100-Hz train, bushy neurons in old mice fire spikes at similar rates as those in young mice, but with less temporal precision, and that both firing probability and temporal precision declined in old mice during the sustained phase of the 100-Hz train.

Action potentials driven by synaptic inputs from the auditory nerve at 400 Hz.

To test the function of the endbulb of Held synapse at a higher rate, the auditory nerve was electrically stimulated at 400 Hz (Fig. 5) to assess evoked responses in postsynaptic bushy neurons. Bushy neurons from young mice fired spikes reliably to the onset phase but less reliably during the sustained phase of the 400-Hz stimulus train (Fig. 5, A and C). In contrast, bushy neurons from old mice not only fired less spikes during the onset phase but also failed to fire spikes to the majority of the stimuli during the sustained phase of the 400-Hz train (Fig. 5, B and D). In 7 out of the 35 bushy neurons tested in old mice, no spikes were evoked at all during the sustained phase of the 400-Hz train, and, in 14 other cells, the firing probability per stimulus was below 0.1. In contrast, only 1 out of 19 bushy cells in young mice failed to fire any spikes, and 2 other cells fired at a rate below 0.1 during the sustained phase of 400-Hz trains (Fig. 5G, right). The action potential firing probability per stimulus was significantly lower in bushy neurons from old mice than those from young mice during both the onset and sustained phases of the 400-Hz train (Fig. 5G; young onset median: 0.91, IQR: 0.15, n = 19; old onset median: 0.46, IQR: 0.54, n = 35; Mann-Whitney test: P = 0.0001; young sustained median: 0.36, IQR: 0.42, n = 19; old sustained median: 0.06, IQR: 0.18, n = 35; Mann-Whitney test: P = 0.0001). Despite the decreased firing probability in bushy neurons from old mice during the onset phase of the 400-Hz train, the vector strength of evoked spikes showed no significant difference between the young and old mice (Fig. 5H, left; young onset median: 0.9202, IQR: 0.0823, n = 19; old onset median: 0.9104, IQR: 0.0736, n = 35; Mann-Whitney test: P = 0.387). During the sustained phase of the 400-Hz train, evoked spikes in bushy neurons from old mice showed substantial temporal jitter (Fig. 5B, bottom right, and Fig. 5, E and F, right), with significantly smaller vector strength (median: 0.8236, IQR: 0.1596, n = 19) than those in young mice (median: 0.9201, IQR: 0.0900, n = 16) (Mann-Whitney test: P = 0.0372; Fig. 5H). Notice that vector strength was not available from 3 bushy cells in young and 16 in old mice because they fired no or too few spikes during the sustained phase of the 400-Hz train (Fig. 5H, right).

In summary, compared with bushy neurons in young mice, bushy neurons in old mice fired spikes at similar rates during the onset phase of low-frequency (100-Hz) inputs but fired significantly less spikes during the sustained phase. The temporal precision of evoked spikes significantly declined throughout the 100-Hz train in bushy neurons from old mice. With high-frequency (400-Hz) auditory nerve inputs, the firing probability in bushy neurons from old mice was profoundly decreased during either the onset or the sustained phase. The temporal precision of evoked spikes did not show a difference during the onset phase between two age groups, but it was significantly decreased in old mice during the sustained phase of the 400-Hz train. Because bushy neurons in old mice can fire relatively as normally as those in young mice when directly activated (Fig. 3), the overall decline in firing probability and temporal precision to auditory nerve inputs seem to reflect defective synaptic drive at the endbulb of Held synapse during ARHL, which is consistent with our previous observation of compromised synaptic transmission in old mice (Xie and Manis 2016). This indicates that auditory nerve terminals in old mice are not able to securely pass auditory information through reliable synaptic transmission to central auditory neurons as those of young mice are. The loss of auditory sensory information at these SGC central terminals due to decreased firing probability and temporal precision in old mice is likely to contribute to the perceptual deficits of ARHL.

DISCUSSION

Central terminals of the SGCs, i.e., the auditory nerve synapses in cochlear nucleus, are the interface between the periphery and central auditory system that gates all of the sound inputs (Nayagam et al. 2011). This study showed that auditory information from the periphery is partially lost in both strength and temporal precision during the transmission at this interface under ARHL. The physiological functions of bushy neurons in old mice appear to be relatively normal in that they are capable of firing spikes at similar rates and timing as those in young mice upon direct current injections, despite changes in their intrinsic membrane properties. Therefore, the decline in both firing probability and temporal precision of auditory nerve-evoked spikes in bushy neurons from old mice is mainly due to defects of the synaptic drive at the endbulb of Held, as we observed in a previous study (Xie and Manis 2016). Given that auditory information conveyed across the synapse between endbulbs and bushy neurons is crucial in processing temporal fine structure of sound (Manis et al. 2011), it is expected that the observed physiological changes in old mice would adversely affect auditory perception during ARHL, which is signified by inability to process auditory temporal fine structure as well as difficulty in performing complex auditory tasks (Anderson et al. 2012; Grose and Mamo 2010; Lorenzi et al. 2006; Plack et al. 2014). Thus, functional decline at the central terminals of the SGCs during aging may also be an essential component of ARHL, in addition to the synaptopathy identified at the peripheral terminals of the SGCs (Sergeyenko et al. 2013; Viana et al. 2015), as well as other upstream pathological changes.

Membrane intrinsic properties of the bushy neurons.

Bushy neurons express a complex combination of voltage-gated ion channels, including sodium channels, low-voltage-activated potassium channels (K_LVA), high-voltage-activated potassium channels, and hyperpolarization-activated cyclic nucleotide-gated channels (HCN) (Cao et al. 2007; Manis and Marx 1991; Rothman and Manis 2003a). Each of these channels contributes to different biophysical properties of bushy neurons in encoding and transmitting the auditory information with fine temporal precision to higher nuclei (Cao et al. 2007; McGinley and Oertel 2006; Rothman and Manis 2003b). Particularly, the K_LVA and HCN channels are open at rest in bushy neurons, with opposing actions that contribute to setting the resting membrane potentials as well as the excitability of auditory neurons (Johnston et al. 2010; Leao et al. 2005; McGinley and Oertel 2006; Oertel et al. 2000; Rothman and Manis 2003b). The observed depolarization in resting membrane potentials of bushy neurons from old mice could be due to a downregulation of K_LVA and/or upregulation of HCN channels during aging, which would contribute to making the neurons more excitable in old mice (Fig. 2F). The enhanced neural excitability of bushy neurons in old mice could also involve structural changes, as well as redistribution of ion channels at the axon initial segments (Kuba 2012; Kuba et al. 2015). Regardless of the underlying mechanisms, the depolarized resting membrane potential and decreased current threshold (Fig. 2, C and F) in bushy neurons from old mice suggest that these neurons can fire in response to weaker auditory nerve synaptic inputs. This may represent an enhanced gain in the central auditory system to compensate for decreased sensory inputs from the auditory nerve under ARHL.

Bushy neurons in old mice showed no change in membrane time constant (τ_m) but increased membrane input resistance (R_m) (Fig. 2, D and E). Because of the relationship τ_m = R_m × C_m, where C_m stands for membrane capacitance, the data indicate that the C_m is decreased in bushy neurons from old mice. It suggests that the cellular membrane area is decreased, which could be due to reduction in the soma size of bushy neurons (Ryugo et al. 2010) in old mice, degeneration of dendritic arbors (Isokawa 1997), or both, during ARHL. It is unclear how these changes affect the processing of auditory information in bushy neurons as well as in targets of bushy neurons during ARHL.

Structural changes of the endbulb of Held during ARHL.

Anatomically, the endbulb of Held develops from small and simple endings to large complex structures with reticulated branches during the early stage of life (Limb and Ryugo 2000; Ryugo and Fekete 1982). After maturation, the structure of normally developed endbulbs seems to stabilize through adulthood (Baker et al. 2010; Ryugo et al. 2006). Hearing impairment can significantly influence endbulb anatomy. The size and complexity of the endbulb terminals were reduced in congenitally deaf white cats (Ryugo et al. 1997; Ryugo et al. 1998), whereas the synaptic vesicle density was decreased and the postsynaptic density was increased in adult cats with prehearing onset deafness (Ryugo et al. 2010). Decrease in the size of synaptic terminals was widely observed in animal models of posthearing onset hearing loss, including in aged Fischer-344 rats (Helfert et al. 2003), DBA/2J mice (McGuire et al. 2015), as well as homozygous shaker-2^−/− and DBA/2 mice (Connelly et al. 2016). Given that CBA/CaJ mice show severe hearing loss at 25–30 mo of age, similar anatomical changes are very likely to occur at the endbulbs in these old mice, which would provide the structural basis for the observed changes in old mice in this study.

Synaptic mechanisms at the endbulb of Held during ARHL.

Previous reports about synaptic mechanisms at the endbulb of Held during aging are rare. The endbulb of Held has been a model synapse for decades and widely studied in slice electrophysiology for mechanisms of synaptic transmission (Manis et al. 2011). However, most studies used juvenile animals up to young adults at the age of 2 mo (Ngodup et al. 2015; Wang and Manis 2005). Our recent study (Xie and Manis 2016) used the voltage-clamp recording technique to assess auditory nerve-evoked excitatory postsynaptic currents in aged bushy neurons, which provided the first clue that synaptic transmission at the endbulb of Held in 2-yr-old CBA/CaJ mice is deteriorated during high rates of activity. Under normal conditions, synaptic transmission at the endbulb is depressed during high-frequency stimulus trains due to its high release probability (Brenowitz and Trussell 2001; Wang and Manis 2008; Yang and Xu-Friedman 2009), which results in reduced postsynaptic currents but preserved timing, since synchronous transmitter release is maintained throughout the trains in young mice (Xie and Manis 2016). In contrast, synaptic transmission at the endbulbs in old mice during sustained high-stimulus-rate trains showed significantly reduced synchronous release and concurrently enhanced asynchronous release. Such shift in release mode from synchronous to asynchronous leads to greater depression of EPSC peak amplitudes during the sustained phase of the trains, predicting that transmission of auditory information across the endbulbs would drop in both signal strength and temporal precision during high-rate activities under ARHL. This functional decline was confirmed in the present study. We also suggested that the underlying mechanisms of ARHL at the endbulbs are associated with changes in intraterminal calcium regulation and/or sensitivity of calcium sensors like synaptotagmins (Bacaj et al. 2013), since increasing intraterminal calcium buffering or decreasing action potential-triggered calcium influx effectively restored synchronous release in old mice (Xie and Manis 2016).

It is worth noting that the compromised synaptic transmission was only observed in responses to stimulus trains at 400 Hz in old mice, whereas synaptic transmission at 100 Hz appeared to be relatively normal (Xie and Manis 2016). In contrast, this study showed functional decline in transmitting spikes at both 100 and 400 Hz in old mice. The disparity at 100 Hz may lie in the age difference of the old mice between these two studies. The present study used old mice at 25–30 mo, whereas the previous study used old mice at 20–26 mo. It is expected that the synaptic transmission at the endbulb of Held continues to deteriorate with age, which contributes to the observed changes at 100 Hz at ages used in this study.

Decreased firing probability in bushy neurons from old mice could also be caused by transmission failures in the auditory nerve fibers before presynaptic action potentials invade the endbulb of Held terminal. Loss of myelination was observed in the auditory nerve (SGC peripheral branch) in aged CBA/CaJ mice and elder humans (Xing et al. 2012). It is conceivable that action potentials can fail during high firing rates in the central branches of the auditory nerve fibers with degenerated myelin, preventing the initiation of synaptic transmission at the endbulbs during ARHL. Such a scenario is unlikely in the present experimental setup because, even when stimuli failed to evoke spikes in bushy neurons during stimulus trains, they always elicited excitatory postsynaptic potentials (Figs. 4B and 5B). This shows that synaptic transmission still occurred at the endbulb, suggesting that action potentials successfully propagated through the auditory nerve from the stimulation site to the endbulbs. However, because bushy cells usually receive inputs from more than one endbulb of Held (Cant and Morest 1979; Cao and Oertel 2010; Nicol and Walmsley 2002; Ryugo and Sento 1991; Spirou et al. 2005), the possibility of transmission failure in the auditory nerve cannot be totally excluded; there could be failure in the fiber connection to one endbulb synapsing with a given bushy cell but not others. It is also possible that the observed decline in firing rate and timing is contributed by reduced number of endbulb of Held terminals (as discussed next) that a target bushy neuron receives in old mice. Presynaptic activity of GABA_B receptors was reported to modulate the firing of bushy neurons in young CBA/CaJ mice (Chanda and Xu-Friedman 2010). It is unclear whether and how GABA_B receptor-mediated signaling changes at the endbulb during aging, which may contribute to the decline in bushy cell firing observed in old mice.

Status of SGCs during ARHL.

Loss of SGCs is a key feature of ARHL (Makary et al. 2011; Otte et al. 1978) that signifies the loss of auditory sensory information to the central system. However, the loss is not substantial in number during normal aging, which is <30% in both CBA/CaJ mice and humans by the end of life (Makary et al. 2011; Sergeyenko et al. 2013; Viana et al. 2015). It is worth noting that the loss of SGCs is not universal across all types, but preferably the SGCs with low spontaneous firing rates (Furman et al. 2013; Schmiedt et al. 1996), which encode suprathreshold sound features, including in difficult listening conditions such as background noise (Costalupes et al. 1984; May et al. 1996). It is hypothesized that the selective loss of these low-spontaneous-rate SGCs can reduce the coding precision of suprathreshold sound and therefore may underlie the auditory neuropathy during noise-induced hearing loss or ARHL (Bharadwaj et al. 2014), in which subjects show relatively normal audiometric thresholds and otoacoustic emissions, but their suprathreshold processing is compromised (Starr et al. 1996; Zeng et al. 2005).

It is unclear what types of SGCs give rise to the endbulb of Held terminals that synapse onto the bushy neurons in old mice recorded in this study. Presumably, they were mostly SGCs with high spontaneous firing rates, which represent about 60% of all SGCs under normal conditions (Liberman 1978), but the proportion may be significantly higher in old mice due to the selective loss of the SGCs with low spontaneous firing rates during ARHL. Therefore, the decline in both firing rate and temporal precision in bushy neurons from old mice during ARHL likely especially reflects the functional deterioration of the surviving SGCs with high spontaneous rates. Thus, this study proposes a new mechanism in explaining the reduced sensory input strength as well as coding precision of suprathreshold sounds during ARHL, namely the deteriorated synaptic transmission at the central terminals of surviving SGCs with high spontaneous firing rates. The observation that bushy neurons in old mice show relatively stable onset responses but deteriorated sustained responses at both low and high stimulus rates (Figs. 4 and 5) may provide an alternative explanation for “hidden hearing loss” or auditory neuropathy during aging, in that detecting the existence of a sound relies more on the onset responses (thus “normal” audiometric thresholds), whereas performing auditory tasks relies more on the sustained responses to process continuous auditory information (thus compromised daily hearing performance, especially under challenging conditions like in a noisy environment).

The finding that ARHL is associated with defective performance of surviving auditory nerve terminals has significant clinical implications. Given that the majority of SGCs survive throughout life (>70% in both CBA/CaJ mice and humans), this study suggests that it is clinically meaningful to explore therapeutic manipulations to restore the normal function of surviving SGCs, in addition to the current efforts in preventing loss of SGCs or restoring SGC populations to normal (Shibata et al. 2011). One applicable example would be to improve the performance of the cochlear implant, which is a major component of current methods for treating hearing loss, including ARHL. The cochlear implant relies on the function of surviving SGCs; however, its performance has reached a technical plateau (O'Leary et al. 2009) that calls for new strategies for further perceptual improvements. This study suggests that one approach is to somehow repair the defective synaptic transmission at the surviving SGC central terminals to ensure auditory information can be securely passed to the central auditory nervous system. Our previous study (Xie and Manis 2016) suggeste d that calcium is abnormally regulated at the endbulb of Held synapse during high-frequency firings in old mice, which leads to compromised synaptic transmission that presumably underlies the decreased firing rate and reduced temporal precision observed in this study. Therefore, one strategy to recover the normal function of the surviving SGCs might be to restore normal calcium signaling at the SGC synaptic terminals.

GRANTS

This work was supported by National Institute on Deafness and Other Communication Disorders Grant R03-DC-013396 (R. Xie).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author.

AUTHOR CONTRIBUTIONS

R.X. conception and design of research; R.X. performed experiments; R.X. analyzed data; R.X. interpreted results of experiments; R.X. prepared figures; R.X. drafted manuscript; R.X. edited and revised manuscript; R.X. approved final version of manuscript.