RTN4RL2 mRNA expression in IHCs and SGNs of the mouse cochlea.

(A) RTN4RL2 is an LRR protein and is anchored to the cell membrane via GPI. In the nervous system RTN4RL2 has been implicated to interact with MAG, versican, BAI (Venkatesh et al., 2005; Bäumer et al., 2014; Wang et al., 2021). (B, C) Exemplary images of RNAscope ISH for RTN4RL2 mRNA (green) in the otic vesicles of E11.5 (B) and E12.5 (C) mice. Scale bar = 50 μm. The developing organ of Corti region is marked with an asterisk and the spiral ganglion is indicated with an arrow. (D, E) Exemplary images of RTN4RL2 RNA expression in hair cells of E16.5 (D) and p1 (E) mice. Hair cells are visualized with anti-Myo7a stainings. Scale bars = 10 μm. (F) Representative images of RNAscope ISH for RTN4RL2 mRNA (green dots) combined with immunostaining for neuron-specific marker βIII-tubulin (grey) in paraffin sections of p40 RTN4RL2+/+ and RTN4RL2-/- cochleae. Scale bars = 20 μm.

Pre- and postsynaptic changes at IHCs of RTN4RL2-/- mice.

(A, B) Maximum intensity projections of representative confocal stacks of IHCs from apical cochlear region of 3-week-old RTN4RL2+/+ (A) and RTN4RL2-/- (B) mice immunolabeled against Vglut3, Homer1, and Ctbp2. Scale bar = 5 μm. Images on the right-hand side are zoomed into the synaptic regions. Scale bar = 2 μm. Some of the putative “orphan” PSDs are marked with the white arrowheads. (C) Number of Ctbp2 positive puncta is not changed in IHCs of RTN4RL2-/- mice (RTN4RL2+/+: 9.9 ± 0.42, SD = 1.62, n = 15, N = 2 vs RTN4RL2-/-: 11.4 ± 0.5, SD = 2.25, n = 20, N = 3; p = 0.06, Mann-Whitney-Wilcoxon test). (D) Ribbon volumes are enlarged in RTN4RL2-/- IHCs (RTN4RL2+/+: 0.16 ± 0.007 μm3, SD = 0.09 μm3, n = 165, N = 2 vs RTN4RL2-/-: 0.21 ± 0.005 μm3, SD = 0.09 μm3, n = 259, N = 3; p < 0.001, Mann-Whitney-Wilcoxon test). (E) Homer1 patches which are juxtaposing presynaptic ribbons show decreased volumes in RTN4RL2-/- IHCs (RTN4RL2+/+: 0.36 ± 0.01 μm3, SD = 0.16 μm3, n = 160, N = 2 vs RTN4RL2-/-: 0.26 ± 0.01 μm3, SD = 0.19 μm3, n = 249, N = 3; p < 0.001, Mann-Whitney-Wilcoxon test). (F) Percentage of Ctbp2 puncta juxtaposing Homer1 is slightly decreased in RTN4RL2-/- mice (RTN4RL2+/+: 96.7 ± 1.45 %, SD = 5.44 %, n = 14, N = 2 vs RTN4RL2-/-: 89.5 ± 2.05 %, SD = 9.18 %, n = 20, N = 3; p = 0.03, Mann-Whitney-Wilcoxon test). Data is presented as mean ± SEM. Box-whisker plots show the median, 25/75 percentiles (box), and 10/90 percentiles (whiskers). Individual data points are overlaid.

Cochlear cell densities are not changed in RTN4RL2-/- mice.

(A) Mid/Basal-modiolar sections labeled for βIII-tubulin (green, neurons) from p40 RTN4RL2+/+ and RTN4RL2-/- mice. Exemplary section for RTN4RL2-/- is the zoomed out image presented in figure 1C. (B) Quantitative analysis shows that the density of SGN cell bodies are similar between RTN4RL2-/- and control cochleae. Data is presented as mean ± SD; N = 7 per group. Scale bar = 10 μm. (C, D) IHC (C) and OHC (D) densities are not affected in P15, 1 month and 2 months old RTN4RL2-/- mice.

Intact number but enlarged size of the ribbons in RTN4RL2-/- IHCs.

(A) Maximum intensity projections of representative IHCs from apical, mid and basal regions of RTN4RL2+/+ (left) and RTN4RL2-/- (right) cochleae of p21-30 mice. Synapses are visualized by staining against Ctbp2/Ribeye (ribbons) and Homer1 (PSDs). Scale bar = 5 μm. (B) The number of the ribbons is not affected along the tonotopic axis in RTN4RL2-/- cochleae. (C) The size of the ribbons is increased in IHCs of both apical and middle turns in RTN4RL2-/- cochleae (p < 0.001, Mann-Whitney-Wilcoxon test). N = 6 animals/genotype. Box-whisker plots show the median, 25/75 percentiles (box) and the range (whiskers). Individual data points are overlaid.

Efferent innervation pattern in RTN4RL2-/- cochleae.

(A) Maximum intensity projections of confocal stacks of outer and inner hair cell regions at the apical region of the cochlea from p21-30 RTN4RL2+/+ (left) and RTN4RL2-/- (right) mice. SGN fibers/terminals and efferent terminals are visualized staining for Na+/K+ ATPase and vesicular acetylcholine transporter (VaChT), respectively. Scale Bar = 10 μm (B) Same as (A) but zoomed into the IHC region. Scale Bar = 5 μm (C) Maximum intensity projections of confocal stacks of a row of apical IHC region from p21 RTN4RL2+/+ (left) and RTN4RL2-/- (right) mice. Presynaptic efferent terminals were stained with an anti-synapsin1/2 antibody. Scale bar = 5 μm.

Reduced GluA2/3 signal juxtaposing presynaptic ribbons in IHCs of RTN4RL2-/- mice.

(A) Maximum intensity projections of representative IHC regions from 1- to 1.5-month- old RTN4RL2+/+ (top) and RTN4RL2-/- (bottom) mouse cochleae immunolabeled against Ctbp2/Ribeye (ribbons) and GluA2/3 (AMPA receptors of PSD). Scale bar = 5 μm. The zoom-in regions marked with the white rectagles are presented on the right-hand side. Scale bar = 2 μm. (B) The number of the GluA2/3 positive puncta is drastically reduced in RTN4RL2-/- mice despite the maintained number of presynaptic ribbons (p < 0.001, Mann-Whitney-Wilcoxon test). (C) Disrupted colocalization of Ctbp2/Ribeye and GluA2 immunofluorescence puncta at IHCs of RTN4RL2-/- mice (p < 0.001, Mann-Whitney-Wilcoxon test). N = 6 animals/genotype. (D) Representative images of RNAscope ISH from p4 mice show maintained expression of Gria2 (red dots) in the SGN somata of RTN4RL2-/- mice. Scale bar = 20 μm.

No apparent GluA4 signal juxtaposing presynaptic ribbons at IHC synapses of RTN4RL2-/- mice.

(A, B) Maximum intensity projections of a few imaging planes from apical IHC synaptic regions of 14-week-old N = 3 RTN4RL2+/+ (A) and N = 3 RTN4RL2-/- (B) mice immunostained against GluA4, GluA2 and RibeyeA. Scale bars = 2 μm. (C) Fluorescence intensities of GluA4 and GluA2 channels were analyzed as the sum of all pixel intensities in approximately 1×1×1 μm³ volumes around the center of mass of RibeyeA immunofluorescent puncta (red boxes, top image). Background GluA4 and GluA2 fluorescence signals were calculated in the same way but outside the ribbons, at random spots inside IHCs (pink boxes, bottom image). Images show the maximum intensity projection of the IHC region in an RTN4RL2-/- mouse. (D) Fluorescence intensities of GluA4 (top) and GluA2 (bottom) channels at ribbons of RTN4RL2-/- mice do not show a significant increase compared to the background (RTN4RL2-/- random) but are significantly decreased compared to RTN4RL2+/+ (Mann-Whitney-Wilcoxon test followed by Bonferroni multiple-comparison correction). Box-whisker plots show the median, 25/75 percentiles (box) and 10/90 percentiles (whiskers). Individual data points are overlaid.

Additional non-synaptically engaged SGN neurites in the cochlea of RTN4RL2-/- mice.

(A) Workflow of SBEM imaging at the mouse apical cochlear region. (B) Example images of neurites beneath IHCs from the RTN4RL2+/+ (left) and RTN4RL2-/- (right) mice. Synaptic ribbons are indicated with red arrows. The regions of interest were magnified from single sections of SBEM datasets (insets). Scale bar = 2 μm. (C) 3D rendering of afferent fiber reconstruction with ribbons (red), showing both synaptic (green, type I SGN with ribbon) and non-synaptic (grey, type I SGN without ribbon) populations in the RTN4RL2-/- mouse. Scale bar = 10 μm. Inset: EM images show the SGN segments on which myelin sheaths start to form. Scale bar 2 = μm. (D) Display of classified radial fibers in the RTN4RL2+/+ (right) and RTN4RL2- /- (left and bottom) animals. All fibers were traced from the habenula perforata (cycles) before classification to avoid bias to terminal types. Scale bar = 10 μm. (E) Percentage of radial fibers with ribbon per bundle (RTN4RL2+/+: 91.41 ± 8.35 %, n = 3 bundles, N = 1 vs RTN4RL2-/-: 74.40 ± 14.58%, n = 6 bundles, N = 2; p = 0.032, unpaired t-test). (F) Percentage of unbranched radial fibers per bundle (RTN4RL2+/+: 94.19 ± 5.04 %, n = 3 bundles, N = 1 vs RTN4RL2-/-: 96.98 ± 3.49 %, n = 6 bundles, N = 2; p = 0.226, unpaired t-test).

Quantification of ribbon number and volume in SBEM reconstructions.

(A) The number of ribbons is not changed in RTN4RL2-/- IHCs (RTN4RL2-/-: 10.3 ± 0.61, SD = 2.1, n = 12, N = 2 vs. RTN4RL2+/+: 10.8 ± 0.83, SD = 2.04, n = 6, N = 1; p = 0.54, Mann-Whitney-Wilcoxon test). (B) Ribbon volumes tend to be larger in RTN4RL2-/- IHCs without reaching statistical significance (RTN4RL2-/-: 0.024 ± 0.001 μm3, SD = 0.01 μm3, n = 125 ribbons in 12 IHCs, N = 2 vs. RTN4RL2+/+: 0.021 ± 0.001 μm3, SD = 0.008 μm3, n = 65 ribbons in 6 IHCs, N = 1; p = 0.06, Mann-Whitney-Wilcoxon test). Box-whisker plots show the median, 25/75 percentiles (box) and the range (whiskers). Individual data points are overlaid.

Shifted operation range of Ca2+ channels but intact exocytosis in IHCs of RTN4RL2- /- mice.

(A) Representative current traces from IHCs of RTN4RL2+/+ (top, black) and RTN4RL2+/+ (bottom, red) evoked by step depolarizations. (B) Average Ca2+ current-voltage relationships (IV curves) in RTN4RL2+/+ and RTN4RL2-/- IHCs. (C) Maximal Ca2+ current amplitude is not changed in RTN4RL2-/- IHCs (RTN4RL2+/+: 210 ± 10.9 pA, SD = 40.9 pA, n = 14, N = 5 vs RTN4RL2-/-: -200 ± 8.19 pA, SD = 40.9 pA, n = 25, N = 8; p = 0.47, Student’s t-test). (D) Fractional activation curves of Ca2+ channels calculated from IVs show depolarized shift in channel activation in RTN4RL2-/- IHCs. (E) Voltage of half maximal activation obtained from Boltzmann fit of the curves from (D) is more positive in RTN4RL2-/- IHCs (RTN4RL2+/+: -29.4 ± 0.66 mV, SD = 2.48 mV, n = 14, N = 5 vs RTN4RL2-/-: -26.6 ± 0.59 mV, SD = 2.94 mV, n = 25, N = 8; p = 0.004, Student’s t-test). (F) Voltage sensitivity (k) is not changed in RTN4RL2-/- IHCs (RTN4RL2+/+: 6.92 ± 0.12 mV, SD = 0.46 mV, n = 14, N = 5 vs RTN4RL2-/-: 6.98 ± 0.1 mV, SD = 0.51 mV, n = 25, N = 8; p = 0.93, Mann-Whitney-Wilcoxon test). (G) Average current traces evoked by 50 ms depolarization to -17 mV (top row) and resulting capacitance response (bottom row) from RTN4RL2+/+ (left, black) and RTN4RL2+/+ (right, red) IHCs. Shaded areas represent ± SEM. (H) Exocytic capacitance change (ΔCm, top) and corresponding Ca2+ charge (Q 2+, bottom) evoked by depolarizations (to -17 mV) of various durations (2, 5, 10, 20, 50, 100 ms). Box-whisker plots show the median, 25/75 percentiles (box) and 10/90 percentiles (whiskers). Individual data points are overlaid.

Depolarized shift of Ca2+ channel activation at single AZs but intact presynaptic organization in RTN4RL2-/- IHCs.

(A) Voltage ramp stimulation protocols (top), evoked whole cell currents (middle) and the presynaptic hotspots of Fluo4-FF fluorescence (bottom) of a representative IHC recording. Black and grey colors represent the two stimulations, one being 5 ms shifted over the other. Images on the right show single imaging planes of representative RTN4RL2+/+ (left) and RTN4RL2-/- (right) IHCs filled with TAMRA conjugated Ctbp2 binding peptide (Ctbp2 bp) and Fluo4-FF Ca2+ dye. Ca2+ hotspots are visualized by subtracting the average of baseline planes from the average of 5 planes during stimulation. Scale bar = 2 μm. (B) Average fluorescence-voltage relationships of Ca2+ influx at single AZ from RTN4RL2+/+ and RTN4RL2-/- IHCs show no difference in the maximal Ca2+ amplitude (Bi; RTN4RL2+/+: 1.6 ± 0.17, SD = 1.2, n = 50 AZs vs RTN4RL2-/-: 1.7 ± 0.13 pA, SD = 1.09, n = 69 AZs; p = 0.24, Mann-Whitney-Wilcoxon test). Shaded areas represent ± SEM. (C) Average fractional activation curves of Ca2+ channels at single AZs show intact voltage sensitivity (Ci; RTN4RL2+/+: 5.79 ± 0.32 mV, SD = 2.04 mV, n = 42 AZs vs RTN4RL2-/-: 6.25 ± 0.22 mV, SD = 1.7 mV, n = 59 AZs; p = 0.06, Mann-Whitney-Wilcoxon test) but depolarized shift of Vhalf (Cii; RTN4RL2+/+: -30 ± 1 mV, SD = 6.5 mV, n = 42 AZs vs RTN4RL2-/-: -25.5 ± 0.98 mV, SD = 7.49 mV, n = 59 AZs; p = 0.002, Student’s t-test) in RTN4RL2-/- IHCs. Shaded areas represent ± SEM. (D) Representative immunolabelings of presynaptic proteins show no apparent mislocalization in RTN4RL2-/- IHCs. Scale bar = 5 μm. Box-whisker plots show the median, 25/75 percentiles (box) and 10/90 percentiles (whiskers). Individual data points are overlaid.

Elevated acoustic thresholds in RTN4RL2-/- mice.

ABR thresholds were measured in response to 4, 8, 16, 32 kHz tone bursts and click stimuli. ABR thresholds of individual animals are shown in open circles on top of the mean ± SEM. Statistical significances are reported as *p < 0.05, ***p < 0.001, Kruskal-Wallis followed by Dunn’s multiple comparison test.

Schematic illustration of the key structural and functional changes in the auditory periphery of RTN4RL2-/- mice.

RTN4RL2-/- mice display enlarged synaptic ribbons and depolarized shift in the activation of presynaptic Ca2+ channels in IHCs, as well as reduced size of PSDs juxtaposing presynaptic ribbons. RTN4RL2 deficiency further leads to a decrease in GluA2-4 AMPA receptor subunits at PSDs and results in a subset of type I SGN neurites that reach the inner spiral bundle but do not engage the IHCs.

List of primary antibodies.