We used male Fmr1 knockout (Fmr1 KO) mice on FVB background (FVB.129-Fmr1^tm1Rbd/J, RRID: IMSR_JAX:008909) and their wild-type (WT) littermates (P5, P14 and P48). The mice were genotyped accordingly to the Jackson Laboratory protocol. Mice were bred in the Animal House of the Faculty of Biology, University of Warsaw. The animals were kept in the laboratory animal facility under a 12-h light/dark cycle with food and water available ad libitum. The animals were treated in accordance with the EU Directive 2010/63/EU for animal experiments and Polish regulations. The mice were anesthetized with inhaled isoflurane (5% in the inhaled air) and euthanized by cervical dislocation.

To obtain inner ears, animals were sacrificed according to approved humane protocol and the inner ears were dissected from the skull. Round and oval windows were opened, and bone over the apical turn removed to allow rapid fixation of 4% paraformaldehyde with 4% saccharose in PBS through the cochlear for 5 min. Next, cochleas were transferred into PBS, to obtain their apical turn. Bone over the middle-ear-facing portion of the cochlear spiral and the tectorial membrane were removed by fine forceps. Following tissue were fixated in 4% paraformaldehyde with 4% saccharose for 40 min. For blocking and permeabilization, cochleas were incubated in PBS with 0.5% Triton X-100 and 5% BSA for 1 h, with gentle shaking. Next, tissue was incubated with primary antibodies at 4 °C for overnight, and secondary antibodies Alexa Fluor (1:1,000; Invitrogen) at room temperature for 2 h. The following antibodies and fluorescent dyes were used: mouse IgG1 Anti-CtBP2 Clone 16/CtBP2 (1:500; BD Transduction Laboratories), mouse IgG2a anti-GluA2 (1:200; Millipore); rabbit anti-MyosinVI (KA-15) (1:1,000, Sigma); mouse IgG2a anti-Bassoon (1:500, Santa Cruz); rabbit anti-FMRP (D14F4) (1:100, Cell Signaling); Phalloidin 405 nm (1:1,000, Abcam), rabbit anti-Tom20 (FL-145) (1:250, Santa Cruz). After immunostaining, cochleas were mounted on microscope slides in ProLong Gold antifade reagent (Invitrogen), and coverslipped.

Z-stacks from selected cochlear regions from each ear were obtained on confocal microscope (Zeiss Axio Imager Z2 LSM 700) with oil-immersion objective 63× and 1× digital zoom. Images were obtained in a 2,048 × 2,048 frame (pixel size = 0.05 μm in x and y). Care was taken to minimize pixel saturation in each image stack. Each stack contained the entire synaptic pole of eight to 11 inner hair cells as viewed from the surface of the organ of Corti. The same microscope settings were used for all sections and all repetitions.

The images were analyzed in the 3D Object Counter¹ application running in the ImageJ/Fiji. Intensity threshold settings were adjusted to identify the vast majority of synaptic densities, while excluding non-synaptic staining. The histogram depicting the frequency of ribbons as a function of the volume of the presynaptic, anti-CtBP2 stained puncta was performed using GraphPad Prism version 7.00 for Windows, GraphPad Software, La Jolla California United States, www.graphpad.com.

To maintain proper structure of hair cells in samples collected from adult mice, we removed the membrane from the round window and the stapes footplate of the oval window, and opened a small hole in the bone at the apex of the cochlea. Next, solution for fixation were gently perfused by syringe and perfusion-fixed on ice over-night.

The whole cochlea was cut out together with the bone capsule and transferred 2.5% glutaraldehyde (Electron Microscopy Sciences), 2% formaldehyde (fresh from paraformaldehyde EMS) in PBS at 4 °C overnight. After washing with PBS, the tissue was incubated in 1% osmium tetroxide and dehydrated in graded aqueous ethanol solutions from 25 to 96% (each for 10 min) followed by 100% ethanol (two lesions each for 10 min). Infiltrated with a mixture of 25% and epoxy resin 30 min on ice, 50% epoxy resin for 2 h, 75 and 100% for 24 h at room temperature Then, the samples in 100% epoxy resin were transferred to a flat embedding mold and placed in an oven at 60 °C for 48 h. Samples were cut on an ultramicrotome (Leica) by diamond Diatome knife (Ted Pella) into 100 μm thick slices. The ribbon was transferred to tungsten coated copper meshes and incubated with 1% uranyl acetate followed by lead aspartate. Electron micrographs were obtained on a transmission electron microscope JEM 1400 (JEOL).

Tissue preparation using this protocol allowed for partial reconstruction of hair cell ultrastructure (Bullen et al., 2015). After fixation (cold 2% paraformaldehyde, 2.5% glutaraldehyde solution in PBS) and decalcification cochleas were transferred to 0.1% tannic acid (EMS, United States) in PBS and incubated in this solution overnight at 4 °C. Samples were processed for serial block face scanning electron microscopy (SBF-SEM) following the method of Deerinck et al. (2010). The samples were trimmed and mounted into pins. Smooth surface of sample to SBF-SEM was obtained by using ultramicrotome (Leica) and diamond knife (Ted Pella). Next, the sample were painted with silver paint (Ted Pella) for enhance conductivity. Samples were imaged by SBF-SEM Sigma VP 3View at the Nencki Institute of Experimental Biology through the technical platform (Laboratory of Imaging Tissue Structure and Function). Scanning parameters: variable pressure 25 Pa, EHT 3 kV, aperture 30um, dwell time 2–5 μs and 7–9 μs. Blocks were cut at 50 nm and 150 nm, and the exposed block face imaged with a pixel size of 3.7–4.2 nm and 10–11 nm for synapse and inner hair cells images, respectively.

3D models for whole cell, cellular organelles such as mitochondria and ribbon synapses in the hair cells of WT and Fmr1 KO mice were prepared using Reconstruct software (Fiala, 2005) http://synapses.clm.utexas.edu/tools/reconstruct/reconstruct.stm.

All the statistical details including statistical tests used and number of biological replicates are described in the Figure Legends. The analyses were performed using Prism 10.4.0 software (GraphPad, San Diego, CA, United States).

We examined the expression of FMRP in the hair cells of mice. Staining was performed on tissue isolated from the cochlea of P5 and adult (P80) WT and Fmr1 KO mice (n = 2 per genotype for P5, n = 3 per genotype for adult mice). We used an anti-CtBP2 antibody as a marker for nuclei and synaptic ribbons in hair cells. Staining with an anti-FMRP antibody confirmed the presence of FMRP in the hair cells (inner and outer hair cells) of both young and adult WT mice (Figure 1).

The structural details of hair cells in the cochlea of Fmr1 KO mice have not been explored; therefore, we began our analysis by examining the structure of inner hair cells (IHCs) using confocal microscopy. We conducted immunofluorescence staining using specific antibodies that mark IHC and their pre- and postsynaptic compartments. The anti-CtBP2 antibody was utilized to stain both the synaptic ribbons and nuclei in the IHCs, while anti-Bassoon was used for the localization of the presynaptic active zone protein Bassoon, and anti-GluA2 for the postsynapse. Additionally, phalloidin was used to visualize the stereocilia, and myosin VI was used to stain the cell body. To observe mitochondria within the hair cells, we used the anti-Tom20 antibody (Figure 2).

The confocal microscopy images revealed the typical morphology of the IHCs in the cochleae of Fmr1KO mice. The stereocilia were situated on the apical part of the cell body, anchored to the cuticular plate (Figures 2A,B), with the nuclei located in the apical half of the cell. Mitochondria were distributed throughout the cell body (Figures 2F–H). The basal part of the IHCs housed multiple synaptic ribbons. Our findings show that the synaptic ribbons in IHCs from Fmr1KO mice are positioned opposite the AMPA glutamate receptors (indicated by GluA2 staining) on the postsynaptic side, within the afferent nerve fibers (Figures 2A,B, 3A–F).

Bassoon plays a crucial role in anchoring ribbons close to the presynaptic membrane. To study whether its level is unchanged in Fmr1KO IHC, we carried out immunostaining using anti-Bassoon and anti-CtBP2 antibodies. We observed that numerous Bassoon immunosignals overlapped with the ribbons in the IHCs of both WT and Fmr1KO mice. Additionally, there were distinct Bassoon signals that did not overlap, which likely originated from efferent nerve terminals (Figures 2D,E).

The structural maturation of ribbon synapses in inner hair cells (IHCs) is a complex process that is tightly regulated during early postnatal development. To investigate how ribbon synapses develop in Fmr1 knockout (KO) IHCs compared to wild-type (WT) mice, we conducted confocal microscopy analyses to assess their volume across three developmental stages, from postnatal day 5 (P5) through P14 and adulthood P48 (Figure 3). These stages encompass both pre- and post-hearing onset periods.

To image the presynaptic ribbons we performed immunostaining using an anti-CtBP2 antibody on tissue isolated from the apical turns of the cochlea at P5, P14, and P48 mice (Figure 3). This approach enabled us to quantify the number of ribbon synapses per IHC and assess the distribution of their average size.

We observed that the number of CtBP2/RIBEYE puncta in IHCs significantly decreased from P5 to P14 (following the onset of hearing) and then stabilized. Although the number of synapses per IHC was comparable between wild-type and Fmr1 KO mice (Figures 4A–C), the average volume of fluorescently labeled presynaptic ribbons was notably smaller in P14Fmr1 KO mice across three litters (Figures 4D–F).

Histograms showing the distribution of average synapse sizes (categorized into arbitrarily defined size classes) revealed a higher prevalence of smaller ribbon synapses in the P14 Fmr1KO (0.04–0.18 μm³), compared to a predominance of larger ribbons in wild-type littermates (0.2–0.6 μm³) (Figures 4E, G–I). Although this difference in ribbon volume was less pronounced in adult mice, smaller synapses remained more prevalent in Fmr1 KO mice at P48 (0.06–0.14 μm³) relative to wild-type controls (0.16–0.44 μm³) (Figure 4F).

Histograms representing the distribution of average synapse volume at P5 indicated that ribbons were comparable between wild-type and Fmr1KO animals at the early developmental stage before hearing onset (Figure 4D).

The ultrastructure of hair cells in Fmr1 KO mice has not been studied before, and the observed delay in the structural development of IHC ribbon synapses may influence their mature appearance. Therefore, to investigate the fine morphology of these cells, we employed two complementary methods based on electron microscopy: transmission electron microscopy (TEM) and serial block face scanning electron microscopy (SBF-SEM). The study was conducted at the developmental stage of P48.

First, we evaluated the overall morphology of the organ of Corti in Fmr1 KO mice, and found no abnormalities in its organization. The sensory cells—inner hair cells (IHCs) and outer hair cells (OHCs)—were properly arranged along the basilar membrane, with the tectorial membrane positioned above them. Both IHCs and OHCs exhibited hair bundles, known as stereocilia (Figures 5A,B).

Next, we investigated the ultrastructure of Fmr1 KO inner hair cells using transmission electron microscopy. Figure 5 presents representative images of IHCs from both Fmr1 KO and wild-type (WT) cochleae. The overall morphology of Fmr1 KO IHCs appeared comparable to that of WT cells, with well-defined stereocilia (Figures 5A,B). The shape of the Fmr1 KO IHCs, as well as the distribution of mitochondria and the nucleus, was similar to that observed in WT IHCs.

We observed typical postsynaptic densities and synaptic structures, including presynaptic dense bodies (ribbons) coated with synaptic vesicles. In some synapses, the ribbon was linked to the postsynaptic membrane via filamentous structures. Mitochondria were located beneath the cuticular plate and surrounding the nucleus, and were often associated with strands of intracellular membrane. The overall morphology of IHCs’ mitochondria qualitatively appeared comparable to that of WT.

While transmission electron microscopy (TEM) provides high-resolution images of inner hair cells and ribbon synapses, it is limited in capturing the overall shape and morphology of entire cells. To address this question, we performed three-dimensional reconstructions of whole IHCs from both wild-type (WT) and Fmr1 KO mice using serial block-face scanning electron microscopy (SBF-SEM).

Based on serial image reconstructions, we extracted specific morphological parameters, such as the mean IHC surface area, which showed no significant differences between the two genotypes (Figure 6D). We further reconstructed key cellular components, including the nucleus, mitochondria, synaptic ribbons, and nerve endings contacting individual IHCs. The mitochondrial distribution appeared uniform, and the localization of the nucleus was similar between Fmr1 KO and WT mice.

Typically, inner hair cells exhibit a distinct flask-like shape, characterized by a constriction in the neck region. However, in the Fmr1KO mouse, the IHC displayed a more fusiform body shape compared to the wild-type control (Figure 6). The nucleus was positioned in the apical half of the cell, while mitochondria were primarily concentrated beneath the cuticular plate and surrounding the nucleus.

Afferent terminals were identified by the presence of a postsynaptic density on the afferent ending, accompanied by a corresponding synaptic ribbon (Figure 6, lower panel). We analyzed the spatial distribution of synaptic structures surrounding the IHCs and performed three-dimensional reconstructions of ribbon synapses. These reconstructions did not reveal any notable dysmorphologies in Fmr1 KO mice compared to wild-type controls (Figure 6).

Quantitative analyses of both the number of ribbon synapses per cell (Figure 6E) and the number of afferent nerve endings contacting each IHC (Figure 6F) showed again no significant differences between WT and KO mice, indicating that the overall ultrastructure of IHCs in the inner ear remains intact in the absence of FMRP.

The quality of the SBF-SEM electron microscopy images was sufficiently good to enable detailed reconstruction of ribbon synapses for both mouse genotypes. For each genotype, we successfully reconstructed six synapses (Figure 7). We measured both the surface area of each synapse and the number of synaptic vesicles associated with it (Figures 7B,C). These analyses revealed no statistically significant differences between Fmr1 KO and wild-type mice.