From hidden hearing loss to supranormal auditory processing by neurotrophin 3-mediated modulation of inner hair cell synapse density

Loss of synapses between spiral ganglion neurons and inner hair cells (IHC synaptopathy), leads to an auditory neuropathy called hidden hearing loss (HHL) characterized by normal auditory thresholds but reduced amplitude of sound-evoked auditory potentials. It has been proposed that synaptopathy and HHL result in poor performance in challenging hearing tasks despite a normal audiogram. However, this has only been tested in animals after exposure to noise or ototoxic drugs, which can cause deficits beyond synaptopathy. Furthermore, the impact of supernumerary synapses on auditory processing has not been evaluated. Here, we studied mice in which IHC synapse counts were increased or decreased by altering neurotrophin 3 (Ntf3) expression in IHC-supporting cells. As we previously showed, postnatal Ntf3 knockdown or overexpression reduces or increases, respectively, IHC synapse density and suprathreshold amplitude of sound-evoked auditory potentials without changing cochlear thresholds. We now show that IHC synapse density does not influence the magnitude of the acoustic startle reflex or its prepulse inhibition. In contrast, gap-prepulse inhibition, a behavioral test for auditory temporal processing, is reduced or enhanced according to Ntf3 expression levels. These results indicate that IHC synaptopathy causes temporal processing deficits predicted in HHL. Furthermore, the improvement in temporal acuity achieved by increasing Ntf3 expression and synapse density suggests a therapeutic strategy for improving hearing in noise for individuals with synaptopathy of various etiologies.

Methods
Animals: All experimental procedures complied with the National Institutes of Health guidelines and were approved by the Institutional Animal Care and Use Committee of the University of Michigan, MI, USA. Cochlear supporting-cell specific Ntf3 knock-down mice (Ntf3 KD) derived from Ntf3flox/flox:Plp1/CreERT mice by Cre-recombination) and Ntf3 overexpressing mice (Ntf3 OE derived from Ntf3stop:Plp1/CreERT mice by Cre-recombination) were generated. Ntf3 KD mice and their controls were maintained on C57BL/6 background. Ntf3 OE and their controls were on FVB/N background. Both male and female mice were included in this study.

Tamoxifen administration: Tamoxifen was injected into the intraperitoneal cavity of P 3-10 Ntf3stop: Plp1/CreERT mice or P1-3 Ntf3flox/flox:Plp1/CreERT mice. A 10 mg/ml solution of tamoxifen was obtained by dissolution in corn oil. Injection was 33 mg/kg for Ntf3stop: Plp1/CreERT mice and 50 mg/kg for Ntf3flox/flox:Plp1/CreERT mice.

Real-time quantitative RT-PCR: Total RNA was isolated from the cortical brain and cochlea samples from 1-month-old mice using RNA extraction kit and Qiazol Reagent (RNeasy mini kit; Qiagen, Germany), and DNase treatment was performed (RNase-free; Qiagen). The complementary DNA was synthesized using iScript cDNA synthesis kit (Bio-Rad, #1708891, USA), according to the manufacturers’ protocol. Quantitative RT-PCR was performed on a CFX-96 Bio-Rad reverse transcription polymerase chain reaction detection system (Hercules, CA, USA) using iTaq Universal SYBR® Green supermix (Bio-Rad, # 172-5121, USA) and primer pairs were synthesized by IDT (Coralville, IA, USA). All samples and standard curves were run in triplicate. Water instead of complementary DNA was used as a negative control. The 10 μl reaction contained 5 μl of SYBR Green supermix, 6 pmol of each forward and reverse primer (0.6 μl), 1.9 μl of nuclease-free water, and 2.5 μl of cDNA sample. The mRNA expression levels in PLP1cre ERT: NT3fl or PLP1cre ERT: NT3 STOP versus their control mice were determined by a comparative cycle threshold (Ct) method and relative gene copy number was calculated as normalized gene expression. Ribosomal protein L19 (RPL19) was used as the housekeeping gene. The following specific oligo primers were used for the target genes: Rpl19, F: 5’ACCTGGATGAGAAGGATGAG 3’; R: 5’ACCTTCAGGTACAGGCTGTG 3’; Ntf3, F 5’GCCCCCTCCCTTATACCTAATG 3’; R: 5’CATAGCGTTTCCTCCGTGGT 3’; Vgf: F: 5’GGTAGCTGAGGACGCAGTGT 3’; R: 5’GTCCAGTGCCTGCAACAGT 3’. Changes in mRNA expression were calculated as relative expression (arbitrary units) respective to the control group for each mouse line.

Immunostaining and synaptic counts: Cochleas from 16-week-old mice were prepared for whole-mount imaging. In brief, the samples were fixed with 4% formaldehyde for 2 hours and then decalcified with 5% ethylenediaminetetraacetic acid (EDTA) for 3-5 days. Cochlear epithelia were micro-dissected into five segments for whole-mount processing. The cochlea segments were permeabilized by freeze–thawing in 30%  and then blocked with 5% normal horse serum for 1 hour. Afterwards, the following primary antibodies were used: 1) CtBP2 to visualize synaptic ribbons (mouse anti-CtBP2 at 1:200, BD Biosciences, catalog # 612044, RRID: AB_399431), 2) GluR2 to visualize postsynaptic receptors (mouse anti-GluR2 at 1:2000, Millipore, catalog # MAB397, RRID: AB_2113875) and 3) Myosin VIIa to visualize IHCs (rabbit anti-Myosin VIIa at 1:200; Proteus Biosciences, catalog # 25-6790, RRID: AB_10015251). Secondary antibodies used were Alexa Fluor 488 conjugated anti-mouse IgG2a (1:1000, Invitrogen), Alexa Fluor 568 conjugated anti-mouse IgG1 (1:1000, Invitrogen), and Alexa Fluor 647 conjugated anti-rabbit (1:1000, Life Technologies). Frequency maps were created using a custom ImageJ plug-in. Images were captured from 5.6 to 45.2 kHz using a Leica SP8 with a 1.4 NA 63x oil immersion objective at 3 digital zoom. Offline image analysis was performed using Amira (Visage Imaging). Quantifications of synapses were done by a blind investigator.

Distortion Product Otoacoustic Emissions (DPOAE) and Auditory Brainstem Responses (ABR): DPOAEs and ABRs were performed. Mice were anaesthetized by i.p. injections of xylazine (20 mg kg−1, i.p.) and ketamine (100 mg kg−1, i.p.). The DPOAEs were induced by two primary tones (f1 and f2) and recorded at (2 × f1)−f2. f1 level was10 dB higher than the f2 level and frequency ratio f2/f1 was 1.2. The ear-canal sound pressure was amplified and averaged at 4 μs intervals. DPOAE thresholds were defined as the f2 level that produced a response 10 dB SPL higher than the noise floor. For ABR measurement, subdermal electrodes were placed (a) at the dorsal midline of the head, (b) behind the left earlobe, and (c) at the base of the tail (for a ground electrode). ABRs were evoked with 5 ms tone pips (0.5 ms rise–fall) delivered to the eardrum. The frequencies of tone pips were 5.6, 8, 11.3, 16, 22.6, 32, and 45.2 kHz, with 15 sound levels from 10 to 80 dB SPL for each frequency. The signals were amplified 10,000 times and filtered through a 0.3 - 3 kHz passband. At each level, an average of 1,024 signals was taken after ‘artifact rejection’. Both recordings were performed using National Instruments input/output boards hardware. Offline analysis was performed using Excel and ABR peak analysis software.

Pre-pulse inhibition (PPI) and gap inhibition of the acoustic startle (GPIAS): Mice were tested in a 10 x 4.5 x 4 cm cage inside a sound-isolation chamber that were placed within a sound-attenuating room. An acoustic sound source was located in the upper part of this chamber. The piezoelectric motion sensor attached to the cage detected the vertical force of the startle reflex. All ASR, PPI, and GPIAS stimuli and responses were generated and recorded with Kinder Scientific Startle Monitor (Kinder Scientific, Poway, CA).

PPI tests were used for assessing sensorimotor gating on 8-15 old week mice, twice a week, each 2 days apart. PPI tests were conducted quiet. The startle stimuli were BBN bursts at 120 dB SPL, 20 ms in duration, 0.1 ms rise-fall times. The prepulse was a narrow band sound centered at 8, 12, 16, 24, and 40 kHz, 50 ms in duration, with 2-ms rise-fall ramps. The PPI test consisted of prepulse trials and startle-only trials, which were delivered alternatively. In prepulse trials, a prepulse ended 50 ms before the startle stimulus. Startle-only trials were similar to the prepulse trials, but no prepulse was delivered. PPI startle ratio is the ratio of the startle magnitude in prepulse trials over the startle magnitude in startle-only trials.

The GPIAS paradigm has been used for measuring auditory temporal processing. Gap inhibition was assessed on 8-15 old week mice, twice a week, 2 days apart. The testing consists of two types of trials, gap trials, and no-gap trials that were delivered alternatively. In both trials, the startle stimulus was 20 ms BBN at 120 dB with 0.1 ms rise/fall times. The startle was preceded either by gaps with varied durations (3-, 5-, 15-, 25-, or 50-ms long) embedded in BBN or by a 50-ms gap embedded in a narrow bandpass background sound centered at 8, 12, 16, 24, and 40 kHz at 65 dB. For gaps with varied durations, the response was normalized to the longest gap (50 ms) on the presumption that the maximum level of inhibition (100%) was elicited at the longest gap. The responses and gap durations were fitted with three-parameter logistic regression models. The gap-detection threshold was defined as the value of the fitted curve that elicited 50% of the maximal inhibition achieved per mouse. Gap-startle ratio is the ratio of the startle magnitude in gap trials over the startle magnitude of the paired no-gap trials.

For each mouse, PPI and gap-startle ratios were averaged from 11 sessions. Each session of PPI and gap-PPI test included 60 pairs of prepulse and startle-only trials for PPI or gap and no-gap trials for gap detection (5 prepulse or background sound frequencies, 12 pairs for each frequency). The interval between trials was randomly varied between 5 and 15s. Each session began with a 2-min acclimatization period in the cage before startle testing began, and all tests were conducted in darkness. Startle-only trials and no-gap trial amplitudes greater than or equal to mean ± 2.5 standard deviations were eliminated. When a trial was eliminated, its paired trial was also eliminated. PPI ratios bigger than 1 or gap-startle ratios bigger than 1.1 were excluded. PPI and gap-startle ratios were calculated as the average with prepulse startle amplitude divided by the mean without prepulse startle amplitude.

Statistical analyses: Analyses were performed using GraphPad Prism 6 (GraphPad Software Inc., La Jolla, CA) and RStudio packages. Data are shown as the mean and standard error of the mean (S.E.M.). The number of replicates (n) is indicated in the results section and figure legends. No explicit power analysis was used to predetermine sample sizes, but our sample sizes are similar to those reported in our previous publications. Statistical differences in auditory physiology (DPOAE threshold, ABR threshold, amplitude, and latency), ribbon synapse counts, behavioral background movement, PPI ratio, gap-startle ratios were analyzed using two-way ANOVA, followed by Bonferroni multiple comparisons test. mRNA expression, ASR amplitude, and gap detection threshold were compared using an unpaired Student’s t-test. Correlations were computed using Pearson’s correlation. The statistical threshold was set to alpha = 0.05.