|Home | About | Journals | Submit | Contact Us | Français|
The cochlear nucleus (CN) receives innervation from auditory and somatosensory structures, which can be identified using vesicular glutamate transporters, VGLUT1 and VGLUT2. VGLUT1 is highly expressed in the magnocellular ventral CN (VCN), which receives auditory nerve inputs. VGLUT2 is predominantly expressed in the granule cell domain (GCD), which receives non-auditory inputs from somatosensory nuclei, including spinal trigeminal nucleus (Sp5) and cuneate nucleus (Cu). Two weeks after unilateral deafening VGLUT1 is significantly decreased in ipsilateral VCN while VGLUT2 is significantly increased in the ipsilateral GCD (Zeng et al., 2009), putatively reflecting decreased inputs from auditory nerve and increased inputs from non-auditory structures in guinea pigs. Here we wished to determine whether the upregulation of VGLUT2 represents increases in the number of somatosensory projections to the CN that are maintained for longer periods of time. Thus we examined concurrent changes in VGLUT levels and somatosensory projections in the CN using immunohistochemistry combined with anterograde tract tracing three and six weeks following unilateral deafening. The data reveal that unilateral deafness leads to increased numbers of VGLUT2-colabeled Sp5 and Cu projections to the ventral and dorsal CN. These findings suggest that Sp5 and Cu play significant and unique roles in cross-modal compensation and that, unlike after shorter term deafness, neurons in the magnocelluar regions also participate in the compensation. The enhanced glutamatergic somatosensory projections to the CN may play a role in neural spontaneous hyperactivity associated with tinnitus.
The cochlear nucleus (CN) receives auditory inputs from type I auditory nerve fibers (ANFs) and non-auditory inputs from several somatosensory nuclei, including the spinal trigeminal nucleus (Sp5), trigeminal ganglion, cuneate nucleus (Cu), vestibular nuclei and contralateral CN (Barker et al., 2012; Bukowska, 2002; Burian and Gstoettner, 1988; Doucet et al., 2009; Haenggeli et al., 2005; Zeng et al., 2011; Zhou and Shore, 2004; Zhou et al., 2010). The non-auditory projections terminate predominantly in the CN granule cell domain (GCD) (Zeng et al., 2011; Zhan and Ryugo, 2007; Zhou et al., 2007), which includes the superficial shell region of ventral CN (VCN) and the fusiform cell layer of dorsal cochlear nucleus (DCN) (Hackney et al., 1990; Mugnaini et al., 1980; Weedman et al., 1996; Zhou and Shore, 2004). These projections give CN the unique ability to integrate information from the auditory nerve and somatosensory system (Shore, 2005) and compensate to cochlear damage (Barker et al., 2012; Dehmel et al., 2012; Kaltenbach, 2007; Shore et al., 2008; Zeng et al., 2009) .
The vesicular glutamate transporter subtypes have different distributions in the brain. VGLUT2 is strongly expressed by neurons in auditory subcortical nuclei, whereas VGLUT1 is dominant in the auditory cortex and at variable levels among auditory subcortical nuclei (Hackett and de la Mothe, 2009; Hackett et al., 2011). In the CN VGLUT1 and VGLUT2 have distinct distributions associated with projections from auditory and somatosensory inputs, respectively (Zhou et al., 2007). VGLUT1 is highly expressed in the magnocellular area of VCN and deep layer of the dorsal cochlear nucleus (DCN3), where ANFs terminate on principal neurons. In contrast, VGLUT2 is expressed mostly in the GCD of the CN, which receives primarily non-auditory inputs from somatosensory nuclei (Zhou et al., 2007). Somatosensory terminals from Sp5, Cu and vestibular nuclei as well as terminals from the contralateral CN colabel with VGLUT2 but rarely VGLUT1, whereas type I ANF terminals exclusively colabel with VGLUT1 (Barker et al., 2012; Zeng et al., 2011; Zhou et al., 2007; Zhou et al., 2010). This pathway-specific feature can be used to quantify changes in auditory and somatosensory inputs after unilateral deafening.
Previously, we demonstrated changes in the relative distributions of VGLUT1 and VGLUT2 in the guinea pig CN after unilateral deafening (Zeng et al., 2009): While VGLUT1 decreased in the ipsilateral VCN two weeks post-deafening, VGLUT2 increased in the GCD that receives non-auditory inputs (Zeng et al., 2009). These changes suggested an enhancement in non-auditory glutamatergic inputs. However, the sources of these increased VGLUT2 levels have not been elucidated. Here, we combined VGLUT immunohistochemistry with anterograde tract tracing to determine whether post-deafening increases in VGLUT2 represented increases in the number of projections from the somatosensory pathways to the CN. The results indicate that following unilateral deafening, an increased number of projections from both Sp5 and Cu provide the basis for the increased VGLUT2 distributions. Furthermore, these increased projections occurred not only in the GCD, but also in magnocellular regions, suggesting a differential redistribution of somatosensory inputs to the CN.
Eighteen female pigmented guinea pigs (250–350 g, Elm Hill Breeding Labs, Chelmsford, MA, USA) with normal Preyer’s reflexes were used in this study. Ten were unilaterally deafened and eight were used as controls. In order to quantify changes in VGLUT isoform levels, three unilaterally deafened animals were allowed to survive for three weeks (3w deafened), and three were allowed to survive for six weeks following the deafening (6w deafened). Two age-matched normal animals were used as control for each group (3w deafened and 6w deafened). An additional eight animals were used to identify the locations of VGLUT- co-labeled terminals following tracer injections in Sp5 or Cu. Of these eight animals, four were unilaterally deafened and four were controls. Two deafened animals received tracer injections in Sp5 and two received injections in Cu two weeks following deafening. These animals were sacrificed after another week to allow for anterograde tracing. Four age-matched control animals received tracer injections in Sp5 (N=2) and Cu (N=2) in the same timeframe as the deafened animals. All animals with tracer injections were allowed to survive for one week before being sacrificed. All procedures were performed in accordance with the NIH guidelines for the use and care of laboratory animals (NIH publication No. 80-23), and guidelines provided by the University Committee on Use and Care of Animals (UCUCA) of the University of Michigan (UCUCA protocol # 08539).
ABRs were measured for all animals before and at the end of the experiment prior to euthanization. ABRs were recorded in an electrically and acoustically shielded chamber (Acoustic Systems, Austin, TX USA). Animals were anesthetized with ketamine (40 mg/kg) and xylazine (10 mg/kg) and body temperature was maintained with a thermostatically controlled heating pad. Sub-dermal recording electrodes were placed at vertex (one cm posterior to bregma), reference (ventral to the pinna on the tested ear) and ground (ventral to the pinna on the contralateral ear) sites.
Digital signals were generated and delivered by Tucker Davis Technologies (TDT, Alachua, FL USA) hardware controlled by a Pentium PC, using the TDT software package SigPlay32. Stimuli were generated using a sample rate of 100 kHz at 16-bit resolution. Equalization to correct for the system response was performed on the digital waveforms in the frequency domain. Acoustic stimuli (15 ms tone bursts, with 1.5 ms rise/fall times) were presented at 10 per second. TDT System II hardware and SigGen/Biosig software (TDT) were used to present the stimulus and record responses. Tones (4, 8, 12 and 16 kHz) were delivered through a Beyer driver (Beyer Dynamic Inc., Farmingdale, NY; aluminum-shielded enclosure made in house), using a speculum placed just inside the tragus. Up to 1024 responses were averaged for each stimulus level. Responses were collected for stimulus levels in 10 dB steps at higher stimulus levels, with additional 5 dB steps near threshold. Thresholds were interpolated between the lowest stimulus level where a response was observed, and 5 dB lower, where no response was observed.
Ten animals were anesthetized with intramuscular injections of ketamine hydrochloride (Ketaset; 40 mg/kg) and xylazine (Rompun; 10 mg/kg) and placed in a stereotaxic frame (David Kopf, Tujunga, CA). Rectal temperature was maintained at 38° + 0.5°C with a thermostatically controlled heating pad. Using a post-auricular approach, a small incision was made behind the left ear and the skin retracted to expose the bulla. A small hole in the bulla was made and kanamycin (20 μl, 30% in normal saline) was injected into the cochlea using a Hamilton syringe, through the round window. After the injection, the skin overlying the bulla was sutured and the animal was allowed to recover.
Two weeks following deafening, two animals received anterograde tracers [10% biotinylated dextran-amine (BDA), MW 10,000, Molecular Probes] in Sp5 and two animals received injections in Cu, while four control animals likewise received identical tracer injections in Sp5 or Cu. The surgical procedure has been described previously (Zeng et al., 2011; Zhou and Shore, 2004). Briefly, an opening just lateral-caudal to lambda was drilled in the occipital bone. A Hamilton microsyringe equipped with a glass micropipette (20-30 μm tip) was positioned into the left Sp5 (3mm lateral to the midline, 2mm caudal to the posterior edge of the transverse sinus, and 10.4mm ventral to the surface of the dura) or the left Cu (1.8 mm lateral to the midline, 3 mm caudal to the posterior edge of the transverse sinus, and 8.8 mm ventral to the surface of the dura). A total volume of 0.1 μl anterograde tracer (BDA) was injected into the Sp5 or Cu. After removing the pipette, the animals were sutured and allowed one week to recover.
Three weeks following deafening, final ABR measurements were performed in all deafened and control animals with tracer injection before being euthanized with FatalPlus (Vortech Pharmaceuticals, Dearborn, MI; 5mg/kg, intraperitoneally). Three deafened and two control animals without tracer injections also received ABRs and were euthanized. The animals were then perfused transcardially with 100 ml of 0.1 M phosphate-buffered saline (PBS; pH 7.4), followed by 400 ml of 4% paraformaldehyde in the same buffer. The brains were isolated and postfixed for 2 hours at 4°C, then transferred into 20% sucrose in 0.1 M phosphate buffer overnight at 4°C. The brainstems were sectioned on a freezing microtome at a thickness of 40 μm; the cochleae were processed for spiral ganglion counts (see below).
Three deafened animals and two control animals survived for six weeks after deafening. Following final ABR measurements, the animals were euthanized and the brainstem and the cochleae processed as above.
All deafened and control animals were processed for spiral ganglion counts. The density of spiral ganglion cells (SGC) in the left (deafened) cochlea was compared with the right side and the control animals.
Differences in the number of spiral ganglion neurons (SGN) were assessed as described previously (Glueckert et al., 2008; Miller et al., 1997; Zappia and Altschuler, 1989). Briefly, cochleae were rapidly removed following vascular fixation. Cochleae were then decalcified in 5% EDTA and processed for embedding into JB-4 Plus, a glycol methacrylate plastic. Sections were cut at 5 μm in a paramodiolar plane such that mid-modiolar sections had six cross-sectional profiles of Rosenthal’s canal. Plastic sections were mounted on slides, stained with Paragon and coverslipped. The twelve most mid-modiolar sections were selected and every third section was used for quantitative assessment to reduce the chance of counting a cell twice. Digital images of the entire most basal cross sectional profile through Rosethals canal were acquired under bright field optics and exported to a Metamorph Image Analysis workstation. An image of the profile through Rosenthal canal was also acquired at a lower magnification and the outline circled to give an area using that could be used to generate a measure for SGN density within that profile. The area of Rosenthal’s canal on every third section was measured and the number of spiral ganglion neuron profiles counted. The criteria for inclusion as a spiral ganglion neuron included having a round cell body (aspect ratio of length to width 3:1) diameter between 14–20 μm, the presence of a nucleus with a diameter between 7–10 μm, and a homogenous cytoplasm. The use of these criteria minimized the potential for over- or underestimating the number of spiral ganglion cells that can occur when counting cell profiles due to, for example, surviving neurons changing their size significantly.
Forty μm transverse sections of the brainstem were cut and mounted in four series on clean glass slides and air-dried. Two serial sections without tracer injections were used to label VGLUT1 or VGLUT2 only, the two other series were saved as back-up. For the sections with tracer injections, one series was used to label BDA and VGLUT1, one was used to label BDA and VGLUT2, the third series was labeled with only BDA for injection location definition and projection puncta counting. The final series was saved for back-up. To define the injection size and location, all sections in the third series containing the injected tracer were photographed using a fluorescent microscope equipped with a digital camera (Leica, DM). The images were imported to Photoshop and the borders of the injection site were traced and cropped. The cropped images were exported to Metamorph for measuring the areas. The volume of each injection site was calculated by summing the injection areas and multiplying by 120 μm, the separation between two contiguous sections (Moore and Kowalchuk, 1988). This analysis yielded two pairs of size and location-matched Sp5 injections (two controls, two three-week deafened) and two pairs of Cu injections (two controls, two three-week deafened).
The VGLUT1 and VGLUT2 immunocytochemistry procedure used here has been described previously (Zeng et al., 2009; Zhou et al., 2007). Briefly, all tissue processing was done at room temperature (20-22°C). Sections were incubated for 30 minutes in a blocking solution containing 1% normal goat serum in 0.1M PBS with 0.1% Triton X-100, PH 7.4, followed by overnight incubation with primary antibodies, VGLUT1 (polyclonal antibody, generated in rabbit, diluted in 1:1000, Synaptic Systems, Germany, Cat. #135 303) or VGLUT2 (polyclonal antibody, generated in rabbit, diluted in 1:1000, Synaptic Systems, Germany, Cat. #135 403). After thoroughly rinsing in PBS, sections were reacted with the secondary antibody (Alexa Fluor 555-conjugated goat anti-rabbit; Invitrogen) for 2 hours. After rinsing, slides were dehydrated in graded ethanol, and coverslipped using micro-cover gel (Micron Diagnostics, USA). Negative controls were conducted on sections that were not treated with either primary or secondary antibodies, resulting in no immunolabeling. VGLUT antibodies were pre- incubated with corresponding synthetic peptides (Strep-Tag® fusion protein containing amino acid residues 456 - 560 of rat VGLUT1, Cat. # 135-3P, or Strep-Tag® fusion protein containing amino acid residues 510 - 582 of rat VGLUT2, Cat. # 135-4P, Synaptic Systems, Germany), resulting in negative immunolabeling. Western blots previously performed (Zhou et al., 2007) demonstrated anti-VGLUT1 antibody with a single band at ~ 60 kDa on protein extracted from both CN and cerebellum and anti-VGLUT2 antibody as a single band at ~65 kDa, corresponding to the molecular weights predicted for VGLUT1 and VGLUT2, respectively. Positive controls for VGLUT were performed in the cerebellar cortex (Hioki et al., 2003; Kaneko et al., 2002; Takamori et al., 2001). To visualize BDA-labeled Sp5 and Cu terminals colabeled with VGLUT, sections were incubated for 2 hours with Cy2 conjugated with streptoavidin (1:300, Jackson ImmunoResearch), followed by immunolabeling with VGLUT1 or VGLUT2.
Sections were examined using a fluorescent microscope equipped with the appropriate filters for Alexa 555 and Cy2 (Leica, DM). Photomicrographs of immunolabeling were imported to Adobe Photoshop for contrast adjustment, then digitized and imported to Metamorph for quantification. The parameters for digitizing photomicrographs were kept constant across all animals (Zeng et al., 2009; Zhou et al., 2007). Images of VGLUT were taken for all subdivisions of the CN, including DCN1, DCN3, GCD (Shell region and DCN2), AVCN and PVCN. Within a slice series, each region of the CN had three images taken in order to create an average puncta density for that subdivision. To focus on changes in non-auditory inputs, the puncta density of the shell region and DCN2 were combined to yield the GCD density. The GCD has been previously defined as the combined shell and fusiform cell layer regions (DCN2)(Zhou et al., 2007), which receive primarily non-auditory inputs. The VGLUT densities in DCN1, DCN3, GCD, AVCN and PVCN in each group were then compared between the control and deafened groups (Zeng et al., 2009).
Quantification was performed as previously described (Zeng et al., 2009; Zhou et al., 2007). Experimenters were blind as to whether the tissue was from normal or deafened animals. VGLUT density was quantified in each of the following CN regions: DCN1, DCN3, GCD, AVCN and PVCN. Sections for AVCN and PVCN were taken from the middle-frequency regions (i.e. halfway between the most dorsal aspect and the most ventral aspect of the VCN). This method thereby encompassed regions affected by the hearing loss. For each side of each animal, three pictures (40X) were taken at equal intervals from caudal to rostral for each of the above selected regions except DCN1 (i.e., one picture from the 25 th percentile , one from the 50th percentile, and one from the 75th percentile, etc). For DCN1, a higher magnification (63X) was used. The photomicrographs were then transferred to Metamorph for automatic quantification. The total number of VGLUT labeled puncta was divided by the chosen area in each photomicrograph to yield the puncta density. No distinction was made between puncta on cell bodies and puncta on dendrites or in the neuropil. Means and standard errors were calculated for the puncta density of VGLUT1 and VGLUT2. The deafened side was compared with the contralateral side and age-matched normal controls.
Under epifluorescence, the labeled Cu and Sp5 terminals in both CN were manually counted on every section of one series in each group. The quantification was based on terminals in both CNs unless otherwise stated. Cu and Sp5 terminal endings were further classified as: 1) mossy fiber-like terminal endings (MF): large irregular swellings (≥2.5 μm), which usually give rise to collaterals; or 2) small boutons (SB): small and round or oval (<2.5 μm), including en passant and terminal boutons (Zeng et al., 2011; Zhou and Shore, 2004). Each ending was defined by size, location in the CN and double labeling with VGLUT1/VGLUT2. Double labeling of Cu and Sp5 terminals with VGLUTs were determined by frequently switching the filters and adjusting the focus of the objective lens. Colocalization was established when the two different labels (red and green) exactly marked the same profile at the same focusing level. In many cases, double-labeled immunofluorescent terminals were also identified by confocal laser microscopy (Olympus 500). The method of identifying double-labeled Cu/Sp5– VGLUT1/VGLUT2 terminals in the CN corresponded well with the confocal findings. For both Cu and Sp5 terminals, there were 4 categories: 1) SB-non-colabeled with VGLUT1/VGLUT2; 2) SB-colabeled with VGLUT1/VGLUT2; 3) MF-non-colabeled with VGLUT1/VGLUT2; 4) MF-colabeled with VGLUT1/VGLUT2. The terminal counts were obtained from the first series of sections and the number of terminal counts was multiplied by 4 to achieve the total terminal counts. Finally, Abecrombie’s correction was used to adjust possible double counting errors : corrected number = count × [section thickness/ (section thickness + terminal size)] (Abercrombie, 1946). For the unilaterally deafened animals, the contralateral and the ipsilateral CNs were compared to each other and to the normal controls. Statistical analyses were performed using one-way ANOVA, t-test or Chi-square, as appropriate. Significance was determined at p< 0.05.
All ten animals with unilateral intra-cochlear injections of kanamycin had profound, unilateral hearing losses as demonstrated with ABR measurements (Figure 1A). No ABR responses were measurable at 90 dB SPL (the limits of our sound system) in the deafened ears. Figure 1A shows the mean ABR thresholds for the deafened animals at four test frequencies (4, 8, 12 and 16 kHz) for the left (ipsilateral, deafened) ears before and after deafening, as well as sixteen normal control ears in the eight normal hearing animals.
Figure 1B shows the mean spiral ganglion densities for the whole cochlea as well as the basal turn at three and six weeks following deafness. The left side (ipsilateral to deafening) for the whole cochlea (F=219.74; df=2, 15; p<0.001) and the basal turn (F=107.89; df=2, 15; p<0.001) shows significantly decreased spiral ganglion density when compared to the control animals and the right ears at both three and six weeks. Figure 1C shows the normal appearance of spiral ganglion cells in one control animal, and Figure 1D shows the significant loss of spiral ganglion cells in one animal after three weeks deafness. Thus, after unilateral deafness (3 and 6 weeks), spiral ganglion densities are reduced significantly compared with the contralateral (right) side and the control groups. Spiral ganglion densities of the contralateral ear were not significantly different between the three and six week groups and the control animals, similar to those previously reported (Miller et al., 2007; Zeng et al., 2009).
1) Along with decreased SG density, VGLUT1 density was significantly decreased in the VCN ipsilateral to the cochlear injections three and six weeks after unilateral deafness.
Compared to both the control and contralateral data, significant decreases in VGLUT1 labeling were seen in the AVCN and PVCN. Figure 2 shows photomicrographs of decreased VGLUT1 density in the ipsilateral AVCN and PVCN in one animal (three weeks after deafening) in comparison with the contralateral side. The photomicrographs in Figure 2A and 2B show strong VGLUT1 labeling in the contralateral AVCN and PVCN, whereas sparse VGLUT1 labeling was observed on the side ipsilateral to the deafening (Figs. 2C and 2D). Compared to both the contralateral CN and normal controls, quantification reveals that ipsilateral VGLUT1 densities were significantly decreased in ipsilateral VCN (AVCN and PVCN) after three (Fig. 3, F=29.51; df=2, 24; p<0.001 for AVCN, F=87.87; df=2, 24; p<0.001 for PVCN) and six weeks of deafness (Fig. 3, F=34.46; df=2, 24; p<0.001 for AVCN, F=53.78; df=2, 24; p<0.001 for PVCN), . No significant change in VGLUT1 was found in other regions of CN, including DCN1 (not shown in the figure).
2) VGLUT2 density was increased in the ipsilateral GCD, which is the primary recipient of non-auditory inputs. In addition, VGLUT2 is also increased in the ipsilateral magnocellular regions of VCN that predominantly receive auditory inputs at both three and six weeks after unilateral deafness.
Although VGLUT2 is intensely expressed in the GCDs of both normal controls and deafened guinea pigs (Zeng et al., 2009; Zhou et al., 2007), it is significantly increased in the ipsilateral GCD of unilaterally-deafened animals starting at two weeks following deafening (Zeng et al., 2009), and shown here also at three and six weeks following unilateral deafening. Figure 4 shows photomicrographs of increased VGLUT2 density in the ipsilateral GCD (Fig. 4A) of one animal (three weeks after deafening) compared with the contralateral side (Fig. 4D). There is sparse VGLUT2 labeling in the contralateral AVCN and PVCN (Figs. 4E and 4F), but stronger VGLUT2 expression in the ipsilateral AVCN and PVCN (Figs. 4B and 4C). Compared with control animals, there is an up regulation of VGLUT2 in the ipsilateral GCD three (F=4.19; df=2, 24; p<0.05) and six weeks (F=4.00; df=2, 24; p<0.05) following deafening. VGLUT2 densities in ipsilateral AVCN and PVCN are significantly increased at three weeks following deafening (Fig. 5; (F=10.32; df=2, 24; p<0.001 for AVCN, F=8.04; df=2, 24; p<0.05 for PVCN). VGLUT2 increases were also seen in the ipsilateral PVCN at six weeks after deafness when compared with control animals (F=4.18; df=2, 24; p<0.05).
Thus, following unilateral deafening, VGLUT2 was significantly increased in regions of the GCD that receive somatosensory inputs and was also increased in the ipsilateral magnocellular region of CN. No significant changes were seen in DCN1 and DCN3.
Four guinea pigs (two controls, two three-week deafened) received BDA injections restricted to Sp5, and four guinea pigs (two controls, two three-week deafened) received BDA injections restricted to Cu. The locations of all eight injection sites are summarized in serial brainstem templates in Figure 6. The injection sites and volumes were well-matched in the deafened and control animals. The first pair of matched Sp5 injections (A, top row) shows a slightly larger injection site in the control animal (volume = 3.56mm3) than the deafened animal (2.81 mm3). The second pair of matched Sp5 injections (B, second row) has a smaller injection size than the first pair, but also shows a slightly larger injection in the control animal (volume = 2.19mm3) than the deafened animal (volume = 1.26mm3). All 4 Sp5 injections are restricted to the Sp5. The two pairs of Cu injections (lower rows, C and D) have very similar injection volumes, with one control animal (C, solid, volume = 0.49mm3) paired with one deafened animal (C, hatched, volume = 0.44mm3), and a second control animal (D, solid, volume = 0.67mm3) paired with another deafened animal (D, hatched, volume = 0.68mm3).
Even though the injection sites and volumes were well-matched in the deafened and control animals, to help further interpret the changes in Sp5 and Cu projections to the CN with hearing loss, the means and standard errors of the means were computed for the number of projection terminals in both deafened and control animals. The results depicted in Figure 7 show little variation in the amount of CN labeling between control and deafened animals. The projection counts are further divided into those colocalized with VGLUT2 and those not colocalized in Figure 8. The results in Figures 7 and and88 reveal that three weeks following unilateral deafening, there was a significant increase in the numbers of both Sp5 and Cu projection terminals in almost every region in the CN compared with control animals. The terminal distributions of Sp5 and Cu projections in control and deafened guinea pigs were all bilateral with ipsilateral dominance, with most of these terminals located predominantly in the GCD, which receives somatosensory inputs. However, the distribution patterns were differentially altered depending on the projection source. In control animals, the total number of Cu projection terminals was considerably less than the number of Sp5 projections (p<0.05, two-tailed t-test). Cu projections were located primarily in the GCD with a minor presence in DCN3 while Sp5 projections were located primarily in the GCD with a broader distribution in DCN3, AVCN and PVCN. In deafened animals, the total numbers of both Cu and Sp5 projections increased, but the total number of Cu projections remained lower than Sp5 projections (p<0.05, two-tailed t-test). Furthermore, the distributions of projection terminals from Cu and Sp5 were changed differently after three weeks unilateral deafness. In spite of a predominant distribution in the GCD, Cu projections were increased in DCN3 and VCN while Sp5 projections were increased in VCN. These data indicate that three weeks following deafening, the CN received increased numbers of projections from both Cu and Sp5. Furthermore, these increases were observed not only in the GCD, which receives somatosensory inputs, but also in the non-GCD regions, DCN3 and VCN (Fig. 7A, B; Fig. 8A, B).
Additionally, the morphological patterns of projection terminals were altered differentially depending on the source (Fig. 8C-F). As shown previously (Zeng et al., 2011), Cu and Sp5 terminal endings are either MFs or SBs. In control animals, the MF-to-SB ratio for Cu terminals was 1:2.3, while the MF-to-SB ratio for Sp5 terminals was 1:9 (Chi square, p<0.001). In deafened guinea pigs, the MF-to-SB ratio of Cu terminals decreased to 1:4 compared with control animals (Chi square, p<0.01), while the ratio for Sp5 terminals increased to 1:5 compared with control animals (Chi square, p<0.001). Thus, there was a greater increase in Cu SBs but a greater increase in Sp5 MFs compared with control animals (Fig 8 C-F).
Sp5 and Cu terminal endings in CN colabel with VGLUT2 but rarely with VGLUT1 (Zeng et al., 2011; Zhou et al., 2007). In the present study, we examined the association of Sp5 and Cu projection terminals with VGLUT2 in deafened guinea pigs compared with control animals (Fig. 8). Three weeks following deafening VGLUT2-colabeled Cu and Sp5 terminals were significantly increased in the CN compared with control animals (p<0.01, two-tailed t-test for Cu; p<0.001, two-tailed t-test for Sp5; Fig 8 A, B). In control animals, most of the VGLUT2-labeled Cu and Sp5 terminals were located in the GCD. In the deafened animals, whileVGLUT2-labeled Cu and Sp5 terminal numbers still predominated and increased in the GCD, they also increased in DCN3 and VCN. Indeed, there were greater increases in VGLUT2-labeled Cu terminals in DCN3 (Chi square, p<0.001) and increased numbers of VGLUT2-labeled Sp5 terminals in VCN (Chi square, p<0.001), than in GCD alone. These changes are consistent with recently reported increased VGLUT2-mediated vestibular inputs to the DCN (Barker et al., 2012) after acoustic overexposure.
The distribution patterns of VGLUT2-labeled of MFs and SBs from Cu and Sp5 were also changed following three weeks deafening. Most MFs colabeled with VGLUT2, and were located predominantly in the GCD in both control and deafened guinea pigs for both Cu and Sp5 terminals (Figs. 8C-F, ,99 and and10).10). While Cu and Sp5 SBs were also GCD dominant, fewer SBs colabeled with VGLUT2 than MFs. In the deafened animals, VGLUT2-labeled Cu and Sp5 MFs retained GCD predominance. VGLUT2-labeled Cu SBs increased in DCN3 (Chi square, p<0.001) (Figs. 8, ,9),9), and VGLUT2-labeled Sp5 SBs increased in the VCN (Chi square, p<0.001) (Figs. 8, ,10).10). VGLUT2-labeled SBs from Sp5 were the main contributors to the increase in VGLUT2 in the ipsilateral VCN following deafness. In addition to an increased number of VGLUT2-labeled SBs, deafened animals exhibited more VGLUT2-labeled MFs than control animals (p<0.001; Figs. 8, ,10).10). While there were fewer Sp5-labeled MFs in the VCN than in the GCD, their appearance in the VCN suggest that a different form of neural processing may occur three weeks after unilateral deafness.
Compared with VGLUT2, there were considerably fewer Cu and Sp5 terminals colabeled with VGLUT1 in both control and deafened guinea pigs (Fig. 11). VGLUT1-labeled Cu and Sp5 endings were GCD predominant in both control and deafened animals. Although VGLUT1-labeled Cu and Sp5 terminals constitute a minor subset of the somatosensory inputs to the CN, VGLUT1-labeled MFs from both Cu and Sp5 were increased in the GCD after deafness, indicating that somatosensory compensation three weeks after unilateral deafness not only involves VGLUT2, but also VGLUT1.
Since type I ANF terminals in the CN exclusively colabel with VGLUT1 (Zhou et al., 2007), the decrease in VGLUT1, together with decreases in spiral ganglion counts, likely indicate decreased ANFs inputs to CN neurons after unilateral deafening. Decreases in the synaptic vesicle marker, SV2 (D’Sa et al., 2007) and synaptophysin, also indicative of synaptic degeneration of ANFs in the CN following cochlear damage (D’Sa et al., 2007; Du et al., 2012; Muly et al., 2002), support this contention. The early decrease in VGLUT1 in the DCN1 post deafening (Zeng et al., 2009) is not maintained three and six weeks post deafening, suggesting a temporary shift of parallel fiber intrinsic connections (Zhou et al., 2007). The lack of change in VGLUT1 in the GCD and DCN3 for up to six weeks after deafening is consistent with previous studies (Benson et al., 1997), although down-regulated synaptophysin was shown in a limited region of the DCN (Du et al., 2012). This may reflect input from a smaller number of ANF synaptic endings in the DCN (Cohen et al., 1972), which receives inputs from multiple sources, including descending, intrinsic, somatosensory pathways, as well as VCN T-stellate cells (Barker et al., 2012; Benson et al., 1997; Haenggeli et al., 2005; Ito et al., 2011; Kane and Conlee, 1979; Ryugo et al., 2003; Weinberg and Rustioni, 1987; Zeng et al., 2011; Zhou et al., 2010). Thus, cochlear damage can lead to synaptic plasticity in the DCN (Du et al., 2012; Hildebrandt et al., 2011; Illing et al., 2005; Kraus et al., 2011; McBain, 2008; Zeng et al., 2009), but may not result in significant down-regulation of VGLUT1 in the DCN.
The major objective of this study was to determine the source/s of the increased VGLUT2 in the CN following unilateral deafness. Several possibilities exist, including intrinsic pathways or synaptogenesis, or extrinsic inputs. Here, we focused on the two major sources of the VGLUT2-mediated inputs to the CN, the somatosensory brainstem nuclei, Sp5 and Cu (Zeng et al., 2011; Zhou et al., 2007). After volume and location-matched tracer injections in Cu or Sp5 in three week-deafened animals, there were significantly more terminal endings in the deafened animals. This increase occurred in spite of slightly larger injections in the normal animals. There was also little variation in the amount of CN labeling between control and deafened animals. The increased VGLUT2 levels, along with increased numbers of VGLUT2-labeled Sp5 and Cu terminals, likely indicate an increase in the number of axonal inputs from these somatosensory nuclei in response to auditory deafferentation. This VGLUT2-mediated compensation begins at two weeks (Zeng et al., 2009) and is maintained up to six weeks (the longest time tested) following unilateral deafness. Other sources such as the vestibular (Barker et al., 2012), lateral reticular (Cui and Shore, 2008; Zhan and Ryugo, 2007) and contralateral cochlear nuclei (Doucet et al., 2009; Zhou et al., 2010) constitute other sources that may contribute to the increase of VGLUT2.
In addition to their upregulation in the GCD, more terminal endings from Sp5 and Cu were found in the ipsilateral magnocellular VCN three and six weeks after unilateral deafness. This redistribution of VGLUT2-labeled non-auditory inputs to regions previously dominated by VGLUT1-labeled ANF inputs has significant implications for signal-processing in the CNs of deafened animals. While physiological studies have suggested that increases in glutamatergic inputs to the DCN may explain deafness-induced hyperactivity and tinnitus (Dehmel et al., 2012), redistribution of these glutamatergic inputs to the VCN opens the way for contributions of the VCN to hyperactivity and tinnitus (Vogler et al., 2011).
The granule cell-rich GCD is the main destination of somatosensory terminal endings (Weedman et al., 1996; Zeng et al., 2011; Zhou and Shore, 2004). MFs, which contain numerous round synaptic vesicles (Mugnaini et al., 1980), play an important role in activating DCN cells via granule cell axons (Ryugo et al., 2003; Weedman et al., 1996; Wright and Ryugo, 1996; Zeng et al., 2009; Zeng et al., 2011). In normal hearing animals, MF terminals from Sp5 and Cu predominantly colabel with VGLUT2 (Zeng et al., 2011). Here, VGLUT2- labeled Sp5 and Cu MFs in the GCD were significantly increased, indicating that VGLUT2- MF-mediated somatosensory input is amplified following three week unilateral deafness. The high density of Na+-channels and fast-gated presynaptic Ca+ channels in MFs can produce a large but brief presynaptic Ca+ influx, triggering transmitter release with high temporal precision (McBain, 2008). Thus, somatosensory MFs may provide a faster-acting pathway after hearing damage. This would be especially evident in the Sp5 to CN pathway since MFs from Sp5 showed a greater increase in the GCD compared with MFs from Cu. These findings provide an anatomical basis for the decreased latencies of Sp5 activation of DCN principal cells after deafness (Shore et al., 2008). One undesirable consequence of the post-trauma neuroplasticity could be increased spontaneous rates in the CN neurons that have been observed after hearing loss (Brozoski et al., 2002; Dehmel et al., 2012; Shore et al., 2008; Sumner et al., 2005). This elevated neural activity in the DCN has been proposed as a physiological correlate of tinnitus (Bauer et al., 2007; Brozoski and Bauer, 2005; Kaltenbach, 2007; Kaltenbach, 2011). Tinnitus indeed, may reflect a predominant increase in somatosensory synaptic strength to the CN (Dehmel et al., 2012). The present data provides a rationale for tinnitus treatment involving blocking the VGLUT2-mediated somatosensory input.
Somatosensory SBs, located in DCN and the VCN shell region (Zeng et al., 2011; Zhou and Shore, 2004), likely mediate slow-acting pathways. Here, auditory deafferentation resulted in more SB terminals from both SP5 and Cu to the CN, but with a different distribution. Like MFs, most SBs from Sp5 and Cu were located in the GCD in normal animals. But after three weeks of unilateral deafness there were more VGLUT2-labeled Sp5 SBs in the VCN and more VGLUT2-labeled Cu SBs in DCN3. This was not seen one and two weeks after deafening (Zeng et al., 2009), suggesting that SBs and MFs are involved longer term effects, but terminate in different regions of the CN. The increased somatosensory SBs in VCN likely activate the principal neurons in the VCN, increasing their role in auditory-somatosensory integration after deafness. The existence of a bushy-cell network in the VCN, for example, may facilitate these interactions since the dendritic inputs on bushy cells are of cochlear as well as non-cochlear origin (Gomez-Nieto and Rubio, 2009).
Synaptic plasticity after cochlear damage is reflected in increased growth associated protein 43 (GAP43) in ipsilateral VCN. Deafness-induced GAP-43-mediated synaptic plasticity in the VCN is predominantly based on cholinergic afferents (Meidinger et al., 2006), and has been proposed to possibly reduce central hyperactivity and tinnitus (Kraus et al., 2011). The increased VGLUT2-labeled Sp5 terminals in the VCN shown in this study indicate that synaptic plasticity in the VCN is also mediated by glutamatergic inputs, which may increase the excitatory drive to the CN and produce tinnitus. It is possible that the cholinergic and glutamatergic amplifications after deafness counteract each other or work synergistically, concepts that require further investigation.
We thank Cameron Rule for his critical reading and editing of this manuscript, and Dr. Richard Altschuler and his lab for the assessment of spiral ganglion neurons. This work was supported by grant R01 DC DC004825 and a core center grant P30 DC-05188 from the National Institute on Deafness and Other Communication Disorders, National Institutes of Health.