Activity-dependent formation of a vesicular inhibitory amino acid transporter gradient in the superior olivary complex of NMRI mice
In the mammalian superior olivary complex (SOC), synaptic inhibition contributes to the processing of binaural sound cues important for sound localization. Previous analyses demonstrated a tonotopic gradient for postsynaptic proteins mediating inhibitory neurotransmission in the lateral superior olive (LSO), a major nucleus of the SOC. To probe, whether a presynaptic molecular gradient exists as well, we investigated immunoreactivity against the vesicular inhibitory amino acid transporter (VIAAT) in the mouse auditory brainstem.
Immunoreactivity against VIAAT revealed a gradient in the LSO and the superior paraolivary nucleus (SPN) of NMRI mice, with high expression in the lateral, low frequency processing limb and low expression in the medial, high frequency processing limb of both nuclei. This orientation is opposite to the previously reported gradient of glycine receptors in the LSO. Other nuclei of the SOC showed a uniform distribution of VIAAT-immunoreactivity. No gradient was observed for the glycine transporter GlyT2 and the neuronal protein NeuN. Formation of the VIAAT gradient was developmentally regulated and occurred around hearing-onset between postnatal days 8 and 16. Congenital deaf Claudin14 −/− mice bred on an NMRI background showed a uniform VIAAT-immunoreactivity in the LSO, whereas cochlear ablation in NMRI mice after hearing-onset did not affect the gradient. Additional analysis of C57Bl6/J, 129/SvJ and CBA/J mice revealed a strain-specific formation of the gradient.
Our results identify an activity-regulated gradient of VIAAT in the SOC of NRMI mice. Its absence in other mouse strains adds a novel layer of strain-specific features in the auditory system, i.e. tonotopic organization of molecular gradients. This calls for caution when comparing data from different mouse strains frequently used in studies involving transgenic animals. The presence of strain-specific differences offers the possibility of genetic mapping to identify molecular factors involved in activity-dependent developmental processes in the auditory system. This would provide an important step forward concerning improved auditory rehabilitation in cases of congenital deafness.
KeywordsActivity-dependent Auditory brainstem Claudin 14 LSO NMRI SPN Tonotopy Deafness VGAT VIAAT
Anteroventral cochlear nucleus complex
Cochlear nucleus complex
Dorsal cochlear nucleus complex
Glycine transporter 2
Interaural level differentiation
Lateral superior olive
Medial nucleus of the trapezoid body
Medial superior olive
Posteroventral cochlear nucleus complex
Region of interest
Superior olivary complex
Superior paraolivary nucleus
Vesicular inhibitory amino acid transporter
During the analysis of various transgenic mouse models [9, 10, 11, 12], we observed occasionally a gradient in immunoreactivity against the vesicular inhibitory amino acid transporter VIAAT (also known as VGAT) in the LSO. This protein serves as an ideal marker for inhibitory synapses, as it is the only transporter filling synaptic vesicles with the inhibitory neurotransmitters glycine and GABA [13, 14, 15]. Here, we investigated VIAAT-immunoreactivity (VIAAT-ir) in a systematic way during development in a total of six different mouse models: four different wildtype-strains, one transgenic mouse line with congenital deafness and one mouse line with cochlear ablation. This design enabled us to identify an activity-dependent and strain-specific formation of a VIAAT gradient in the SOC.
Animals and tissue preparation
Overview of mouse lines and number of animals used
P48 (cochlea ablated)
Bilateral cochlea ablation
Cochlea ablation was performed as described previously . NMRI mice aged P27 were anesthetized with a mixture of Ketamine and Xylazine (90 and 7 mg/kg bodyweight i.p., respectively). The head of the animal was fixed in a stereotaxic instrument and the pinnae of both ears were tied to the front. By a deep cut behind the ear, the eardrum was uncovered and ruptured with a hollow needle to expose the auditory ossicles. After detaching the malleus, the whole organ of Corti was removed using a hollow curved needle. The wound was cleaned, filled with gel foam and surgically closed. Operated animals were administered the analgesic Novalgin (200 mg/kg bodyweight i.p.). Subsequently, animals were housed in separated cages. Weight and health conditions were checked daily. Operation was considered successful, if operated animals showed no Preyer reflex in response to sound exposure. Animals were perfused as described above 3 weeks after cochlea removal.
Free floating immunohistochemistry was performed as described previously [10, 18]. Briefly, slices were incubated in blocking solution containing 3% bovine serum albumin, 10% goat serum and 0.3% Triton in PBS (pH 7.4) for 1 h at room temperature. Primary antibodies rabbit anti-VIAAT (1:500; RRID: AB_887871; Catalog No. 131002, Synaptic Systems, Göttingen, Germany), rabbit anti glycine transporter 2 (GlyT2) (1:200; RRID: AB_2619997; Catalog No. 272003, Synaptic Systems) and mouse anti-NeuN (1:1000; RRID: AB_10711040; Catalog No. ab104224, Abcam, Cambridge, UK) were diluted in carrier solution (0.3% Triton, 1% BSA, 1% goat serum in PBS, pH 7.4). Slices were incubated with primary antibody solution over night at 7 °C, rinsed in PBS and incubated in carrier solution containing an Alexa Fluor coupled secondary antibody (Alexa 488 goat anti-rabbit or Alexa 594 goat anti-mouse, 1:750, Invitrogen, Carlsbad, CA, USA) for 1.5 h at room temperature. Subsequently, slices were again rinsed in PBS, mounted on slides, air dried and cover slipped. Although, specificity of the anti-VIAAT antibody was previously shown in VIAAT knockout animals by the manufacturer (https://www.sysy.com/products/vgat/facts-131002.php), control experiments were performed by blocking the primary antibody with the corresponding control peptide (Catalog No. 131-OP, Synaptic Systems). This synthetic peptide (AEPPVEGDIHYQR) has been used for immunization. Only a faint immunosignal in the MNTB was visible after this treatment (data not shown). Specificity of NeuN and GlyT2 antibodies were tested by the manufacturer in immunohistochemistry and immunoblotting. For all primary antibodies used omission of secondary antibodies resulted in the absence of immunosignal (data not shown).
Full focus image stacks of the SOC were taken with a BZ 8100 E fluorescence microscope (Keyence, Neu-Isenburg, Germany), equipped with Nikon Plan Apo 4× and 20× objectives. A constant exposure time was set for all images. Files were obtained using the Keyence BZ observation software and processed in Adobe Photoshop CS6 (Adobe Systems) or Image J (U. S. National Institutes of Health, Bethesda, MD, USA). To be able to count VIAAT-positive synaptic boutons, detailed images of the LSO were taken with a TCS SP2 confocal microscope (Leica, Wetzlar, Germany), equipped with a Leica HCX PL APO 63× oil immersion objective. Image stacks of 1024 × 1024 pixels were taken with fixed parameters concerning laser power and pinhole settings. Images were processed using Image J.
Gray value analysis
To analyze VIAAT-ir in the lateral and medial limb of the LSO and SPN, regions of interests (ROIs) of 70 × 70 μm or 50 × 50 µm, respectively, were analyzed for their mean gray values (3 animals per age, strain and genotype, 4–6 LSO sections per animal). Background fluorescence was determined in a third ROI outside of the SOC in a non-nucleic region. The mean value of the medial limb was set to 100%. Bars depict the relative lateral gray value compared to medial (dashed line). Using SPSS Version 23.0 (IBM Corp., Armonk, NY, USA) the data set was defined as non-Gaussian distributed. Statistical analysis was therefore performed using the Mann–Whitney U test. Error bars illustrate the standard deviation (SD). Analysis was carried out blind to the respective genotype.
VIAAT-positive boutons were counted in a 70 × 70 µm ROI of confocal image stacks of the lateral and medial limb of the LSO. To determine the number of VIAAT-immunoreactive boutons, images were converted into binary images and the watershed routine was applied to separate particles. Single particles were considered as boutons between a size of 0.44 and 2.20 µm as defined previously . Larger particles were expected to be accumulations of single boutons. To obtain the total number of boutons, the size of accumulations was divided by 1.3 (i.e. the average bouton size). Error bars depict the SD. Student`s t test was used as a test for statistical significance after testing for Gaussian distribution. Analysis was carried out blind to the respective genotype.
Developmental VIAAT-gradient in the LSO of NMRI mice
The auditory system of mice is not fully developed at birth and hearing-onset occurs around P10 . We therefore investigated the time course of gradient formation. Since distinction of lateral and medial regions in the SPN is not as clear as in the LSO, most of our analysis focused on the LSO. Developmental analysis revealed that the VIAAT gradient in the LSO is mainly established between P8 and P16 (Fig. 2). While there was no visible gradient of VIAAT-ir in the LSO at P8, i.e. prior hearing-onset (Fig. 2A c, c′), it was clearly visible at P16, which is about 4 days after hearing-onset (Fig. 2A b, b′). The gradient became even more pronounced at P30 (Fig. 2A a, a′). Quantification of VIAAT-immunosignal confirmed this observation (Fig. 2C). Gray values of the lateral limb were 22% higher in P16 animals (medial: 1.453 ± 0.158, lateral: 1.770 ± 0.198; p = 0.001) and 31% at P30 (medial: 1.298 ± 0.123, lateral: 1.702 ± 0.204, p = 0.000012) compared to the medial limb. Thus, the VIAAT-ir gradient forms during postnatal development around hearing-onset. Additional quantification of the VIAAT-ir in lateral and medial regions of the SPN at P30 revealed a 36% higher VIAAT-ir in lateral regions of this nucleus (Fig. 2B, medial: 1.925 ± 0.379, lateral: 2.579 ± 0.346, p = 0.000071).
Activity-dependent formation of the VIAAT-gradient in the NMRI strain
To investigate whether also maintenance of the gradient is activity-dependent, we analyzed mice with cochlear ablation after hearing-onset. To this end, NMRI mice were subjected to bilateral cochlea removal at P27, i.e. roughly 2 weeks after hearing-onset. Successfully operated animals were killed 3 weeks after deafening at P48. Subsequent anti-VIAAT immunohistochemistry revealed the persistence of the gradient in these mice (Fig. 5A b, b′). There was a significant 18% upregulation of the immunosignal in the lateral limb compared to the medial one (Fig. 5B; medial: 1.176 ± 0.115, lateral: 1.379 ± 0.122. p = 0.000242). This result indicates that activity deprivation after hearing-onset does not influence the VIAAT gradient, at least during a 3-week period. Taken together, these results reveal that neuronal activity is required for the formation of the VIAAT-ir gradient in the LSO, but not for its maintenance.
No VIAAT-gradient in other wildtype mouse strains
Here, we report a gradient of VIAAT-ir in the LSO and SPN of NMRI mice with high expression in the lateral limb and lower expression in the medial limb. This VIAAT-gradient does not result from an increased bouton number in the lateral limb compared to the medial one as shown by quantification of boutons in the LSO (Fig. 3). The constant number of VIAAT-positive boutons in both limbs of the LSO in NMRI mice suggests a higher expression of VIAAT in single boutons of the medial limb. This can be explained by either an increased number of transporters per vesicle or by a larger vesicle pool in lateral boutons.
Tonotopic organization of the auditory system
Overall, the gradient followed the medio-lateral axis of the LSO and SPN and thus their tonotopic organization as low frequencies are processed in the lateral and high frequencies in the medial part of these nuclei. Tonotopic organization, i.e. the tuning of topologically neighboring neurons to neighboring frequencies, is a characteristic property of the auditory system (Fig. 1). Matching this tonotopic organization, several molecular gradients in expression patterns have been observed in the auditory system. In the cochlea, there is an opposing gradient of Kcnq4-expression, coding for the voltage gated potassium channel Kv7.4, in inner and outer hair cells [25, 26]. Additionally, the calcium binding proteins calbindin and β-parvalbumin display opposing gradients along the cochlea [27, 28]. On the level of the auditory brainstem, gradients of the potassium channels Kv1.1 and Kv1.3 are present in MNTB and LSO [29, 30, 31]. In the LSO of rats and gerbils there is a medio-lateral gradient of the glycine receptor with a higher abundancy in the medial limb [32, 33, 34]. Additionally, stronger immunoreactivity against the inhibitory neurotransmitters glycine and GABA was found in the medial limb of the LSO in gerbils, although this is most likely attributed to a higher neuron number in this limb . So far, no gradient in the SPN was reported.
Presumably, these gradients provide the molecular basis for an efficient transduction of high frequencies in the medial limb of the auditory nuclei as fast repolarization and a reliable inhibition is crucial for the computation of high frequency sounds [36, 37]. Of note, the here reported VIAAT gradient shows strongest labeling in the lateral limb, which contrasts the previously described gradients of proteins associated with inhibitory transmission in gerbils and rats, which show high labeling of glycine receptors in the medial limb [32, 33, 34]. The physiological relevance of these inverse oriented gradients has still to be elucidated.
Activity-dependent formation of tonotopic gradients
The formation of tonotopic gradients often depend on the interplay between molecular factors and neuronal activity [38, 39]. Our analysis revealed that neuronal activity is required for the formation of this gradient, as the gradient is disrupted in deaf Cldn14 −/− mice. The appearance of the gradient between P8 and P16 indicates that its formation requires sound-driven activity. Furthermore, spontaneous activity is likely present in the Cldn14 −/− mouse model, as hair cells appear normal up to P7, and rapid loss of outer hair cells begins only thereafter . However, a recent analysis of mice lacking the α9 subunit of nicotinic acetylcholine receptors demonstrated that already subtle alterations in the temporal fine structure of spontaneous activity disrupt proper development of the auditory brainstem such as correct axonal pruning . To distinguish between the role of spontaneous versus sensory evoked activity in formation of the gradient, we attempted cochlear ablations around hearing-onset. However, initial experiments yielded poor survival rates. Thus, further studies are required to firmly conclude that the gradient depends solely on sensory-evoked activity.
Activity-dependent formation of tonotopic features agrees well with a previous analysis of tonotopic features in MNTB neurons of wildtype or congenital deaf dn/dn mice. Electrophysiological and molecular investigations in wildtype mice revealed gradients in high-and low threshold potassium conductances as well as in immunoreactivity for Kv3.1 and Kv1.1, contributing to these conductances. These tonotopic gradients were abrogated in congenital deaf dn/dn mice . A similar observation was made for HNC4 channels and the associated hyperpolarization-activated currents . Furthermore, the non-uniform distribution of glycine receptors in the LSO of gerbils is not established in cochlea ablated animals . Finally, a recent analysis revealed an activity-dependent formation of a neuronal size gradient of MNTB neurons .
Sensory activity is not only required for proper formation but also for maintenance of molecular patterns. A genome wide microarray analysis performed 3 or 21 days after cochlear ablation in adult rats detected 449 or 1.031 genes, respectively, with altered expression pattern. Among them were many genes encoding proteins for neurotransmission such as SNAP25 or GABA receptors . As this study indicated stronger impact 3 weeks after sensory deprivation, we probed the role of activity for maintenance 21 days after cochlear ablation. This experimental design, however, did not interfere with the presence of the gradient. Therefore, sensory activity is apparently not required for the maintenance of the VIAAT-gradient once it has been established.
Strain-specific formation of the VIAAT-ir gradient
Intriguingly, the gradient was only present in NMRI mice among four different mouse strains tested (Fig. 2, 6). Noteworthy, this is to our knowledge the first time of a mouse strain-specific feature in tonotopicity. Well-established strain-specific auditory features refer to age-related hearing loss or differences in auditory brainstem responses [44, 45, 46]. A splice variant of cadherin 23 which causes early onset of hearing loss after ~ 13 weeks of age is only present in a subset of mouse strains [47, 48]. Cadherin 23 was shown to be involved in tip-link formation between stereocilia of cochlear hair cells . With respect to the central auditory pathway, detailed analysis of auditory brainstem responses revealed variation between 10 tested mouse strains. This concerned the amplitude increase with intensity, which reached a maximum and then leveled off or decreased with higher intensity, inter-peak I–IV latencies or the slope of latency–intensity function . Unfortunately, NMRI mice had not been included in this study. A previous comparison of hearing thresholds in C57Bl/6J and NMRI mice at 7 month of age showed good hearing ability in the NMRI strain . The good hearing of NMRI mice is also reflected in the frequent use of this strain in behavioral auditory studies [51, 52, 53, 54, 55]. However, to our knowledge there is no known characteristic in the hearing ability of NMRI mice that could account for the observed VIAAT-gradient in the LSO and SPN of these animals.
Finally, strain-specific formation of the VIAAT-gradient will provide a powerful tool for the identification of genetic factors involved in activity-dependent processes in the auditory system. This is of great importance in the auditory field, as deafness is among the most common sensory disorders [56, 57]. Cochlear implants are available to restore peripheral dysfunction, but do not bypass central auditory defects, which frequently occur .
One possibility is quantitative trait locus mapping . This technique was already successfully used to identify a mutation in cadherin 23 associated with age-related hearing loss . Recent advances such as the availability of chromosome substitution strains further alleviate this approach . Alternatively, comparative transcriptome analysis can be performed between the SOC of NMRI mice and C57Bl6/J to identify genetic factors. A similar approach had recently identified crystallin N as an important protein for proper auditory brainstem development [12, 62]. A strain-based approach will be superior to the comparison of normal hearing and deaf mice, as the differences in gene expression are likely subtler between strains, facilitating the identification of candidate genes.
In this study, we reported a higher VIAAT-ir in lateral regions of the LSO and SPN compared to medial regions. The establishment of this gradient is developmentally regulated and forms between P8 and P16 in NMRI mice. Analysis of other commonly used wildtype strains revealed no such gradient, suggesting a strain-specific characteristic. This specificity paves the way to identify genetic factors involved in activity-dependent processes in the auditory hindbrain. This will be important with respect to auditory rehabilitation. Congenital deafness will interrupt those processes, therefore impairing the benefit from cochlear implants. Currently, the molecular mechanisms governing activity-dependent processes in the auditory system are unknown, thereby preventing new therapeutic approaches for restoring central auditory defects. Furthermore, our findings extend the auditory repertoire of strain-specific features, thus sending a note of caution when comparing data from different mouse strains.
This study was designed by HGN and LE. LE analyzed 129/SvJ, C57Bl/6J and Cldn14 −/− mice, the bouton number in NMRI mice and the GlyT2 and NeuN immunohistochemistry. MW did the immunohistochemistry for NMRI and CBA/J mice and performed the cochlea ablations. Interpretation of the data was done by LE and HGN. LE designed Fig. 1. LE and HGN wrote the manuscript. All authors read and approved the final manuscript.
We thank Martina Reents for expert help with genotyping and Benjamin Illing for initial help with the cochlea ablation.
The authors declare that they have no competing interests.
Availability of data and materials
The datasets analyzed during this study are available from the corresponding author on request.
Consent for publication
Animals used in this study were bred in the animal facility of the University of Oldenburg in accordance with the current German Animal Protection Law. All protocols were approved by the regional animal care and use committee (LAVES, Oldenburg, Germany) and followed the NIH Guide for the Care and Use of Laboratory Animals.
This work was supported by the DFG grant No428/18-1 and the Cluster of Excellence EXC 1077 Hearing4all.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 36.Caspary DM. Focus: GABA and glycine neurotransmission in mouse auditory brainstem structures. Boca Raton: CRC Press LLC; 2001. p. 317–20.Google Scholar
- 42.Weatherstone JH, Kopp-Scheinpflug C, Pilati N, Wang Y, Forsythe ID, Rubel EW, et al. Maintenance of neuronal size gradient in MNTB requires sound-evoked activit. J. Neurophysiol. 2016;jn.00528.2016.Google Scholar
- 50.Heffner RS, Koay G, Heffner HE. Focus: Sound localization acuity changes with age in C57Bl/6J mice. 2001. p. 31–35.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.