Animals and ethical statement

34 female Long Evans rats (9−10 weeks, 200–300 g) originating from Charles River (St Germain sur l’Arbresle, France) were used for the experiments. All experiments were performed in accordance with the European Union 2010/63/EU Directive and under veterinary and National Ethical Committee supervision (French Agriculture Ministry Project Authorization: APAFIS #40187−2022120518126271). The present study was approved by Neurosciences Ethic Committee N°71 of the French National Committee for animal experimentation. All personnel participating in this study were trained in animal welfare and care and possessed the mandatory French certification for animal experimentation. All surgeries were performed under isoflurane or ketamine and medetomidine anesthesia depending on the surgery (see either chemical unilateral labyrinthectomy section and auditory brainstem response test section). Humane endpoints—based on grooming, natural behavior, hydration status, clinical signs such as breathing/ventilation, response to handling, and anorexia (> 20% loss of pre-operative body weight)—triggered immediate euthanasia. No animal in this study reached these endpoints, and all animals were perfused at the end of the experiment for immunohistochemical analysis. Every attempt was made to minimize both the number and suffering of the animals used in the experiment. The animals were housed at the Centre de Recherche en Psychologie et Neurosiences (Aix Marseille University) animal facility, with 12 h – 12 h diurnal light variations and free access to water and food.

Study design

For vestibular assessment, 28 rats were assigned to two groups (SHAM, n = 12 and cUL, n = 16) for behavioral and immunohistochemical assessment and were manipulated for 14 days prior to the ototoxic exposure. Among them, one animal was excluded from the D30 cUL group (see Criteria for exclusion section). Thus, 23 female rats (SHAM, n = 12 and cUL, n = 11) were used for behavioral assessment and killed 30 days after cUL for immunohistochemical analysis (see below). To monitor vestibular dysfunction, a qualitative assessment was conducted at a preoperative session and daily for the first 10 days post-cUL, and subsequently on D15, D21, and D30, to define the acute, post-acute, and compensated phases of the vestibular syndrome. Based on the time course of vestibular deficits observed in this qualitative assessment, time points for the postural assessment (Dynamic Weight Bearing 2) [33] and light/dark box test (LDBT) were selected to capture key stages of behavioral recovery. These tests were therefore performed at preop, D1, D4, D7, D9, D15, D21, and D30 postop.

For immunohistochemical analysis, the peripheral vestibular system (Scarpa’s ganglion and utricle), cochlea and central vestibular nuclei were collected at D3 (D3 cUL, n = 4) and D30 (SHAM, n = 5 and D30 cUL, n = 4). All cUL animals on D3 received an injection of BrdU (bromodeoxyuridine, 200 mg/kg, i.p) and were killed either 3 h after the injection (n = 4) or 30 days after surgery (n = 4) to study acute cell proliferation and long-term plasticity. All SHAM animals on D3 received the same BrdU injection but was killed only 30 days after surgery (n = 5) (Fig 1).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Study design used to assess the behavioral and cellular effects of chemical unilateral labyrinthectomy (cUL).

In total n = 12 animals received a SHAM and n = 16 rats a cUL surgery. The peripheral vestibular system (Scarpa’s ganglion and utricle), cochlea and central vestibular nuclei (medical vestibular nuclei, MVN) were studied at D3 (n = 4, D3 cUL group) and D30 (n = 5, SHAM group and n = 4, cUL group) time points. cUL group on D3 received an injection of BrdU (bromodeoxyuridine, 200 mg/kg, i.p) and was killed either 3 h after the injection (n = 4) or 30 days after surgery (n = 4) to study acute cell proliferation and long-term plasticity. SHAM group on D3 received the same BrdU injection but was killed only 30 days after surgery (n = 5). Quantitative assessment of vestibular syndrome and anxiety-like behavior were performed one day before the surgery and 1, 4, 7, 9, 15, 21, 30 days after the surgery in SHAM (n = 12) and cUL (n = 11) group. s: sacrifice.

https://doi.org/10.1371/journal.pone.0355839.g001

For auditory assessment, 6 additional rats (SHAM, n = 2 and cUL, n = 4) were used to analyze auditory bainstem responses (ABR) at preop, D3 and D30 postop.

Chemical unilateral labyrinthectomy

cUL group (n = 20) was subjected to a left-side chemical unilateral labyrinthectomy, following the surgical procedure previously reported in the literature [10,19,34,35]. Thirty minutes after a buprenorphine (Buprecare®, 0.03 mg/kg, s.c.) injection, the rats were anesthetized with isoflurane (induction 5% for 4 minutes, mask 2.5%). After local anesthesia with 1% lidocaine (0.1 mL, s.c.), a left paramedian incision was made to expose the lamboidal ridge and the external ear canal. The external ear canal was opened just anterior to the exit point of the facial nerve. With homemade forceps, the tympanic membrane was perforated caudally to the hammer shaft, and 150 µL of a 20% bupivacaine solution was instilled into the tympanic cavity with a 26-gauge needle. During about 3 min, the bupivacaine solution was aspirated and instilled slowly again. After the local anesthetic was instilled, the same procedure was performed to instill 150 µL of a 10% solution of p-arsanilic acid during about 4 min. After the last thorough aspiration, the wound was closed by 3 stitches (vicryl 3−0). Before awakening, a solution of Ringer Lactate (Virbac, 10 mL/kg, s.c.) was injected to alleviate the dehydration resulting from the rats’ inability to drink normally due to the ototoxic exposure. Post-operatively, animals were monitored daily for 1 week, corresponding to the critical phase of the vestibular syndrome, and high-fat food was provided in the cage to compensate for food intake and support recovery. Body weight and general condition were assessed daily using a standardized clinical scoring grid throughout this period. The success of the procedure was attested upon awakening by the onset of characteristic vestibular symptoms. In the SHAM surgery group (n = 12), 150 µL of 0.9% saline was instilled into the tympanic cavity during 7 min.

Criteria for exclusion

Animals were excluded from the study if no typical vestibular syndrome was observed at awakening or if the following symptoms were observed:

• loss of body weight equal to more than 20% of the pre-treatment,
• ulcer of the cornea, which could occur due to an inadvertent lesion of the facial nerve,
• bleeding from the tympanic cavity, which could prevent the diffusion of bupivacaine or p-arsanilic acid into the inner ear,
• abnormality in behavioral qualitative scoring, e.g., convulsions, paresis, or hemiataxia.

Following these criteria, one animal has been excluded from the D30 cUL group because no typical vestibular syndrome was observed at awakening, resulting in n = 11 animals in the D30 cUL group included in the study for behavioral analysis.

Auditory brainstem responses (ABR)

The hearing thresholds before, 3 days, and 30 days after surgery were assessed using auditory brainstem responses (ABR). The ABR recordings were collected using needle electrodes positioned under the skin, at the skull vertex (active electrode), behind left or right mastoid (reference electrode), and in the neck muscle (ground electrode). Signals were amplified 10⁴ times, filtered between 300 and 3000 Hz (Grass ICP 511 amplifiers), digitally converted, and averaged with a Micro1401 Plus system (Cambridge Electronic Devices, UK). Auditory stimuli were tone pips (with 2ms linear rise/fall time with no plateau) at octave frequencies from 2 to 32 kHz at a repetition rate of 10 s^-1 presented via an in-ear miniature earphone placed in one ear at a time. Tone pip level varied from 90 dB SPL to 0 dB SPL using 10-dB steps. For each ear, stimulus repetitions incremented from 200 at 90 dB SPL to 1800 at 0 dB SPL. ABR were analyzed offline using a custom-written MATLAB program. The five ABR wave maxima correspond to the different brainstem nuclei along the auditory pathway starting from the auditory nerve (wave I) to the inferior colliculus (wave V). The threshold was estimated as the lowest intensity of stimulation that yielded a repetitive waveform. Threshold shifts were evaluated by subtracting the ABR threshold obtained before surgery from the threshold obtained 3 and 30 days after surgery.

Behavioral assessment

Cellular investigations

Statistical analysis

We performed all statistical analysis using GraphPad Prism software (version 10, GraphPad Software). Summary graphs are all shown as mean ± SEM. A two-way repeated measures ANOVA with two post-hoc tests were used for behavioral data. Sidak’s was used to determine statistical differences between the groups and Dunnett’s multiple comparison to compare values at each postoperative time with the preoperative value (delay effect). A two-way repeated measures ANOVA with Tukey’s post-hoc tests was used for the threshold shift analysis. Sphericity Greenhouse-Geisser correction has been applied to the data set when the normality was not verified. The statistical analyses of cellular data in the deafferented MVN were evaluated by nested one-way ANOVA followed by post hoc analysis with Tukey’s test. Kruskal Wallis’ test and Dunn’s multiple comparison post hoc test were used for calyx counting in the utricle. To determine the number of animals used for the behavioral assessment in this study, we conducted a power analysis with the G × Power software. We set the effect size at 0.25, the alpha at 0.05 and the power (1-beta) at 0.95 with a group number of 2 and 8 measurements throughout time. The power analysis gives us a total sample size of 24 and therefore, we used 24 rats in total for the behavioral assessment, but one rat was excluded (see section 5.4). p values of < 0.05 were considered statistically significant. * p < 0.05, ** p < 0.01, *** p < 0.001.

Chemical unilateral labyrinthectomy induces long term peripheral vestibular alteration

The peripheral vestibular system detects environmental stimuli notably through the utricle, an otolith organ which contains type I hair cells surrounded by calyces (Fig 2A). These calyces relay information to the vestibular nuclei via the Scarpa’s ganglion. Calyces were entirely lost on the side exposed to p-arsanilic acid (left) as early as D3 post-cUL and remained permanently absent at D30 post-cUL, compared to SHAM (*** p < 0.001 for both time points, Fig 2B). Unexpectedly, a transient decrease in calyx number was also detected on the side contralateral to ototoxic exposure at D3 post-cUL (** p < 0.01 vs SHAM and *** p < 0.001 vs D30 cUL, Fig 2C). Unlike the utricle, the Scarpa’s ganglion on the side ipsilateral to ototoxic exposure remains intact after chemical unilateral labyrinthectomy (Fig 3) with cell bodies without swelling, regardless of the time after this ototoxic exposure. In addition, representative Myo7a immunostaining of contralateral cochlea confirmed the preservation of hair cells, supporting the absence of detectable contralateral cochlear damage following cUL (S1 Fig). In conclusion, cUL induced a complete and permanent loss of type I hair cells afferents in the utricle exposed to p-arsanilic acid, accompanied by adaptive plasticity in the contralateral utricle relative to ototoxic exposure, while Scarpa’s ganglion remained unaffected.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Chemical unilateral labyrinthectomy permanently destroys calyces in the utricle ipsilateral to ototoxic exposure.

A. Confocal immunostaining images of NKAα3+ (red) and DAPI (blue) in the utricles ipsilateral (left) and contralateral (right) to ototoxic exposure for SHAM group (n = 5), three days (D3 cUL, n = 4) and thirty days (D30 cUL, n = 4) after cUL. Scale bar = 10 µm. Quantitative assessment of the cUL effect on the number of NKAα3 + calyces in the ipsilateral utricle (B) and in the contralateral utricle (C) relative to ototoxic exposure. The histogram represents group means ± SEM, and each point corresponds to one analyzed z-stack image, with three images analyzed per animal. A significant difference from the SHAM group is indicated by * in blue for D3 cUL and D30 cUL groups. A significant difference between the D3 cUL and D30 cUL group is indicated by * in black (** p < 0.01, *** p < 0.001; Kruskal Wallis’ and Dunn’s multiple comparison post hoc tests).

https://doi.org/10.1371/journal.pone.0355839.g002

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Chemical unilateral labyrinthectomy preserves Scarpa’s ganglion.

Illustration of methylen blue in the left Scarpa’s ganglion of the SHAM (n = 5) and D30 cUL (n = 4) groups. Scale bar = 50 µm.

https://doi.org/10.1371/journal.pone.0355839.g003

Chemical unilateral labyrinthectomy alters cochlea structure and auditory function on the side ipsilateral to ototoxic exposure

An overview of the cochlea (x10), along with magnified views at x20 and x63, is presented in Fig 4A. Structural damage to the cochlea is observed from D3 and persists at D30, on the side ipsilateral to ototoxic exposure. These alterations affect hair cell afferents (NKAα3+) and nuclei (DAPI) compared with the SHAM group. Although no cell quantification or formal scoring was performed for this descriptive observation, the ipsilateral cochlea consistently appears altered relative to SHAM group. In contrast, the side contralateral to ototoxic exposure remains intact at both D3 cUL and D30 cUL, similar to the SHAM group (S2 Fig). Auditory brainstem responses (ABR) were assessed preoperatively and at 3 and 30 days after the surgery in cUL (n = 4) and SHAM (n = 2) groups (S3 Fig). Threshold shift analysis revealed a severe hearing loss in the ipsilateral ear (left) relative to ototoxic exposure in cUL animals at both postoperative time points (*** p < 0.001 vs. preoperative ABR threshold; Fig 4B), with an average loss of approximately 70 dB. No significant change in auditory thresholds was detected in the contralateral ear (right) in cUL animals relative to ototoxic exposure (Fig 4C). In the SHAM group, tympanic membrane perforation appeared to induce a mild hearing loss of about 30 dB, restricted to the left ear (Fig 4D and 4E).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Chemical unilateral labyrinthectomy alters the cochlea structure and audition solely on the side ipsilateral to ototoxic exposure.

Orthogonal maximum intensity projection of confocal immunostaining images of the cochlea on the side ipsilateral to ototoxic exposure (left), showing NKAα3+ (red) labeling of hair cells afferents and DAPI (blue) staining, in SHAM group (n = 5), three days (D3 cUL, n = 4) and thirty days (D30 cUL, n = 4) after cUL. Arrows indicate the thin layer of hair cells afferents present in the SHAM group and absent in the cUL group in acute (D3) or compensated (D30) delays. Scale bar = 50 µm. Curves indicate the mean auditory brainstem responses (ABR) threshold shifts in the ipsilateral (left, B) and in the contralateral (right, C) ears relative to ototoxic exposure in cUL animals (n = 4), as well as in the ipsilateral (D) and contralateral (E) ears relative to the surgery in SHAM animals (n = 2). Error bars represent SEM. A significant difference from the pre-operative value is indicated by * in blue. A significant difference between cUL D3 and cUL D30 is indicated with * in black. Statistical significance was only assessed in the cUL group using two-way ANOVA followed by Tukey’s multiple comparison post hoc test (** p < 0.01, *** p < 0.001).

https://doi.org/10.1371/journal.pone.0355839.g004

Chemical unilateral labyrinthectomy induces central medial vestibular nuclei plasticity

Unilateral vestibular loss induces numerous plastic events in the vestibular nuclei [46]. Among these events are cell proliferation, glial reaction, and neural KCC2 modulation, which are all involved in vestibular compensation [23,42,47]. We evaluated cell proliferation with a single injection of BrdU (200 mg/kg) 3 h prior to intracardiac perfusion at D3 post-cUL (S4A Fig). Only about 2.95 ± 0.8 BrdU+ cells are found in the MVN (S4B Fig) compared to about 20 BrdU+ cells after unilateral vestibular neurectomy [36,48,49]. In our chemical labyrinthectomy model, no survival hair cells were found at D30 post-cUL (n.s vs SHAM group).

Reduced KCC2 expression at the neuronal membrane reflects elevated intracellular Cl- ion concentration, weakening GABAergic inhibition and potentially shifting GABAergic signaling to excitatory [50]. In our model, there is no decrease in KCC2 membrane expression (S5A-B Figs). Unlike in the unilateral vestibular neurectomy, few cell proliferations, no newly generated survival cells, nor any neuronal KCC2 modulation have been found in this cUL model.

cUL induced glia cell plasticity in the MVN (Fig 5A) with a significant increase in microglia (IBA1; F_(2,10) = 7.146; ** p < 0.01 vs SHAM group), astrocytes (GFAP; F_(2,10) = 13.47; ** p < 0.01 vs SHAM group) and oligodendrocytes (Olig2; F_(2,10) = 35.96; *** p < 0.001 vs SHAM group) at D3 post-cUL in the MVN (Figs 5B, 5D and 5F). These results are consistent with those previously described in UVN model [24,36]. At D30 post-cUL, only a significant increase in astrocytes persists (F_(2,10) = 13.47; * p < 0.05 vs SHAM group) although microglial number remains high and shows no difference with D3 cUL group. cUL might induce morphological changes in microglial cells (Fig 5C). The fluorescence IBA1 + area increases at D3 cUL (F_(2,10) = 45.03; *** p < 0.001 vs SHAM group) and decreases at D30 cUL (F_(2,10) = 45.03; * p < 0.05 vs SHAM group). Since the number of IBA1 + cells in D30 cUL is similar to that in SHAM with a tendency toward an increase in microglia, the marked decrease in fluorescence might reflect a smaller immunoreactive area and an amoeboid morphology of these cells. However, consistent with an increased number of GFAP+ cells in the MVN at D3 and D30 after cUL, the fluorescence GFAP+ area (Fig 5E) increases as well compared to SHAM group (F_(2,10) = 15.41; ** p < 0.01).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Chemical unilateral labyrinthectomy modulates glial reaction in the deafferented medial vestibular nucleus.

A. Confocal immunostaining images of IBA1+ , GFAP+ and Olig2+ cells (green) in the deafferented medial vestibular nucleus (MVN) for SHAM group (n = 5), three days (D3 cUL, n = 4) and thirty days (D30 cUL, n = 4) after UL. Scale bar = 50 µm. Quantitative assessment of the cUL effect on the number of IBA1+ (B), GFAP+ (D), Olig2+ (F) cells in the deafferented MVN of SHAM (grey), D3 cUL (blue) and D30 cUL (blue) groups. Quantitative assessment of the cUL effect on the fluorescence IBA1+ (C) and GFAP+ (E) area in the deafferented MVN of SHAM (grey), D3 cUL (blue) and D30 cUL (blue) groups. The histogram represents group means ± SEM, and each point corresponds to the mean number of cells per animal analyzed. A significant difference from the SHAM group is indicated by * in blue for D3 cUL and D30 cUL groups. A significant difference between the D3 cUL and D30 cUL group is indicated by * in black (* p < 0.05, ** p < 0.01, *** p < 0.001; nested one-way ANOVA and Tukey’s post hoc test).

https://doi.org/10.1371/journal.pone.0355839.g005

Chemical unilateral labyrinthectomy induces typical vestibular disorders

The behavioral deficits were assessed at several time points over a period of 30 days using a qualitative vestibular score (see Methods section 5.5.1.). Vestibular syndrome was observed as early as D1 post-cUL (8.73 ± 0.5), with a peak at D4 (12.9 ± 0.69), reflecting the progressive development of the ototoxic lesion induced by cUL (Fig 6). Thereafter, the vestibular deficits decreased through two distinct phases. Until D9, during the acute phase, vestibular symptoms remained pronounced, with only a limited reduction in syndrome severity despite the onset of functional recovery. Then, in the post-acute phase, a progressive and linear reduction in vestibular score was observed, reaching a value of 4.36 ± 0.21 at D15 (Fig 6). In the compensated phase, from D15 to D30, the vestibular score reached a plateau, and inter-individual variability decreased, reflecting more similarities in the behavioral patterns across animals. However, the syndrome did not completely disappear 30 days after the cUL (3.45 ± 0.13), with residual signs such as head tilt, bobbing, tail-hanging reflex and locomotion impairment, while barrel rolling and circling behaviors were no longer observed.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Qualitative assessment of vestibular syndrome.

Curve illustrating the time course of the vestibular syndrome of the cUL group divided into three phases: acute (after surgery to D9), post-acute (D9 to D15) and compensated (D15 to D30).

https://doi.org/10.1371/journal.pone.0355839.g006

Chemical unilateral labyrinthectomy induces acute and persistent postural deficit

The support surface area parameter, delimited by the animal’s four paws, is commonly used to describe postural stability and restoration of balance [33,41,51]. The SHAM group maintained a stable support surface area throughout the entire period (21.65 ± 0.88 cm²; no delay effect). For the cUL group, the support surface area was significantly increased as soon as the first post-cUL day (25.55 ± 1.96 cm², Time x Group: F_{(7, 147)} = 5.215; * p < 0.05 vs SHAM group) and did not return to pre-operative values at D30 (27.02 ± 0.89 cm², Time x Group: F_{(7, 147)} = 5.215; *** p < 0.001 delay effect; Fig 7A).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. Effect of the chemical unilateral labyrinthectomy on static postural parameters.

Curves indicating the mean post-operative recovery of the support surface area (A), the body sway (B), the time spent in the front left paw (C) and the weight distribution on the lateral axis (D) of SHAM (n = 12) and cUL (n = 11) groups. Error bars represent SEM. A significant difference from the pre-operative value is indicated by * in black for cUL group. A significant difference between SHAM and cUL group is indicated with * in blue (* p < 0.05, ** p < 0.01, *** p < 0.001; repeated measure two-way ANOVA and Sidak’s post hoc for statistical differences between the two groups and Dunnett’s multiple comparison post hoc test to compare values at each postoperative time with the preoperative value).

https://doi.org/10.1371/journal.pone.0355839.g007

The body sway, an index of instability commonly used in clinical posturology, quantifies the barycenter dispersion while the animal is static and on all four paws (Fig 7B). The SHAM group maintained a stable support surface area throughout the entire period (0.42 ± 0.034 cm²; no delay effect). For the cUL group, the body sway was significantly increased as soon as the first post-cUL day (0.77 ± 0.12 cm², Time x Group: F_{(7, 147)} = 4.607; *** p < 0.001 vs SHAM group) and persisted at D30 (0.84 ± 0.10 cm², Time x Group: F_{(7, 147)} = 4.607; *** p < 0.001 vs SHAM group and delay effect).

SHAM group consistently spent approximately 32.24 ± 2.53% of the time on their front left paw during the 5-minute acquisition regardless of the delay (no delay effect, Fig 7C). In contrast, cUL animals showed increased and sustained use of their front left (ipsilateral side relative to ototoxic exposure) paw from the first post-cUL day to D30 post-cUL (Time x Group: F_{(7, 147)} = 2.748; *** p < 0.001 vs SHAM group at both time points).

We then quantified the percentage of weight distribution along the lateral body axis, between the left and right paws (Fig 7D). SHAM group distributed their weight equally between the left and right paws regardless of the delay (51.38 ± 0.87% vs 48.62 ± 0.87%; no delay effect). Prior to the ototoxic exposure, cUL rats exhibited a balanced weight distribution between the left and right paws (51.18 ± 0.72% vs 48.82 ± 0.72%). These results are consistent with those previously described in UVN model [33,37]. From D4 to D30 post-cUL, cUL rats exhibited a significant increase in their weight applied on the left (ipsilateral side relative to ototoxic exposure) paw (about 57.31 ± 1.03%; Time x Group: F_{(7, 147)} = 2.929; ** p < 0.01 at D4, D21, D30 and *** p < 0.001 at D7-D15 vs SHAM group). In conclusion, cUL induced postural instability and both temporal and weight distribution biases toward the side ipsilateral to ototoxic exposure.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. Effect of the chemical unilateral labyrinthectomy on dynamic postural parameters.

Curves indicating the mean post-operative recovery of the number of left (A) and right (B) circling, and the time spent rearing (C) of SHAM (n = 12) and cUL (n = 11) groups. Error bars represent SEM. A significant difference from the pre-operative value is indicated by * in black for cUL group. A significant difference between SHAM and cUL group is indicated with * in blue (* p < 0.05, ** p < 0.01, *** p < 0.001; repeated measure two-way ANOVA and Sidak’s post hoc for statistical differences between the two groups and Dunnett’s multiple comparison post hoc test to compare values at each post-operative time with the pre-operative value. Sphericity Greenhouse-Geisser correction was applied to the time spent in rearing).

https://doi.org/10.1371/journal.pone.0355839.g008

Chemical unilateral labyrinthectomy induces acute and persistent dynamic deficit

Before surgery, none of the two groups performed left-hand circling (1.17 ± 0.24 for SHAM group and 1.37 ± 0.2 for D30 cUL group; Fig 8A). The SHAM group did not perform left nor right circling after surgery: the number of turns per acquisition varied very little from the preoperative session to the D30 delay. For the cUL group, the number of left fast laps per session increased from D1 post-cUL (12.01 ± 4.36; Time x Group: F_{(7, 147)} = 3.096; * p < 0.05 delay effect) to D30 post-cUL (15.73 ± 5.37; Time x Group: F_{(7, 147)} = 3.096; ** p < 0.01 delay effect) with a peak a D7 (21.82 ± 6.30; Time x Group: F_{(7, 147)} = 3.096; *** p < 0.001 vs SHAM group and delay effect). At the same delay, the cUL group exhibit only about 4.09 ± 0.79 right circling (Fig 8B).

The SHAM group spent about 41.06 ± 3.33% of their time in rearing regardless of the delay after the surgery. In contrast, the cUL group spent only 13.74 ± 3.03% (Time x Group: F_{(7, 147)} = 3.624; *** p < 0.001 vs SHAM group and delay effect, Fig 8C) of their time in rearing, indicating an important instability of the animal in D1 post-cUL. This instability remains at D30 post-cUL (Time x Group: F_{(7, 147)} = 3.624; ** p < 0.01 vs SHAM group and delay effect).

Chemical unilateral labyrinthectomy induces acute and persistent anxiety-like behavior

The LDBT can be used to assess the animals’ level of anxiety-like behavior, as the light zone is naturally avoided by rodents. First, we quantified the number of entries into the light compartment during the whole 5-minutes acquisition (Fig 9A). The SHAM group did approximatively 8.1 ± 0.6 entries into the light compartment regardless of the delay (no delay effect). In contrast, the cUL group first decreased the number of entries from D1 to D4 post-cUL (minimum at 3.18 ± 0.7 entries; n.s. vs SHAM group). This slight decrease can be explained by the increased immobility of the animals in this test because of the vestibular syndrome. Then, the number of entries significantly increased from D9 (16.55 ± 2.4; Time x Group: F_{(6, 108)} = 11.88; ** p < 0.01 vs SHAM group) to D30 post-cUL (18.18 ± 3.65; Time x Group: F_{(6, 108)} = 11.88; *** p < 0.001 vs SHAM group), reflecting the hyperexcitability of injured animals. As for the time spent in the light compartment, the SHAM group maintained a high level of exploration throughout the study with approximately 129.06 ± 8.23 seconds (Fig 9B). In contrast, the cUL group spent significantly less time in the light compartment in acute phase (D1, 66.73 ± 11.24 seconds; Time x Group: F_{(6, 108)} = 2.574; ** p < 0.01 vs SHAM group) and compensated phase (D30, 67.73 ± 16.56 seconds; Time x Group: F_{(6, 108)} = 2.574; *** p < 0.001 vs SHAM group) demonstrating anxiety-like behavior induced by the cUL.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 9. Effect of the chemical unilateral labyrinthectomy on anxiety-like behavior.

Curves indicating the mean post-operative recovery of the number of entries into the light compartment (A) and the time spent in the light compartment (B) in the light/dark box test of SHAM (n = 12) and cUL (n = 11) groups. Error bars represent SEM. A significant difference from the pre-operative value is indicated by * in black for cUL group. A significant difference between SHAM and cUL group is indicated with * in blue (* p < 0.05, ** p < 0.01, *** p < 0.001; repeated measure two-way ANOVA and Sidak’s post hoc for statistical differences between the two groups and Dunnett’s multiple comparison post hoc test to compare values at each post-operative time with the pre-operative value).

https://doi.org/10.1371/journal.pone.0355839.g009

Chemical unilateral labyrinthectomy alters the inner ear

Chemical unilateral labyrinthectomy induces deafferented medial vestibular nuclei plasticity

Our cUL model induced few cell proliferation but no survival, suggesting that this reactional plasticity is specific to vestibular nerve section models [24,89]. This difference may stem from the greater severity of nerve section [90], including loss of the Scarpa’s ganglion [7,89,91]. Other responses, such as potassium-chloride co-transporter KCC2 downregulation in neurons, also seem restricted to nerve section models [42,49]. Our results contrast with those of [92] who reported BrdU+ cells after p-arsanilic acid exposure, but their multiple BrdU injections between D11–D15 likely captured a broader window than our single BrdU injection during the critical period, potentially explaining the discrepancy.

Following unilateral vestibular deafferentation, restoring a level of homeostatic excitability within VN is essential for functional recovery. In the cUL model, this central compensation may rely more on pre-existing neuronal or glial cells than on the generation of new cells. Indeed neuroglia (astrocyte, microglia and oligodendrocytes) are key players responsible for the homeostatic support of the central nervous system [93] and are upregulated in other vestibular models [4,36,94]. Moreover, glial reactivity has been reported in vivo and ex vivo in the deafferented VN and might accelerate vestibular compensation [10]. Our focus will be limited to the deafferented side, given that glial plasticity has been consistently reported to occur only in the deafferented VN [10,49,94].

Astrogliosis, characterized by increased GFAP expression [95,96], was observed in our model at both acute and compensated stages, consistent with typical astrocyte reactivity following injury. Since few cell proliferations has been observed in our model, increased number of GFAP+ cells might result from astrocyte migration or upregulation of GFAP expression in astrocytes that do not express detectable levels of GFAP in sham group. This phenomenon has been previously reported in cerebral pathologies [97]. Beyond this reactive profile, astrocytes are known to support synaptic plasticity and neural network reorganization in the central nervous system [98,99]. In particular, a specific astrocytic subtype emerged in the deafferented MVN, potentially contributing to synaptogenesis [100]. In addition, astrocytes can modulate glutamatergic transmission [101] which is a critical driver for vestibular compensation [102]. As vestibular recovery involves sensorimotor relearning [103], astrocytes, as key component of the tripartite synapse, likely play a pivotal role in the central plastic changes [104], necessary for compensation. Taken together, these findings highlight astrocytes as central regulators of neuronal excitability, positioning them as key mediators of the adaptive plasticity underlying vestibular compensation.

Microglia in reactive states increase IBA1 expression [105], as observed in our model only in acute phase of the vestibular syndrome. Microglia are known to regulate neuronal excitability, by reducing the activity of overly excited neurons, most likely through astrocytes [106,107]. This raises the question of whether they could increase neuronal activity and thus restore excitability in the deafferented VN through other mechanisms. Furthermore, microglia support myelin repair processes by facilitating oligodendrocyte progenitor proliferation and differentiation [108]. Morphologically, homeostatic microglia exhibits a ramified shape, whereas reactive microglia is often associated with hypertrophic forms [109–111]. In our model, microglia morphology seems to differ between acute and compensated phases with amoeboid or deramified form at D30. Since microglia can be protective during the acute phase of injury but become detrimental if reactivity persists, further morphological profiling, such as through Sholl analysis, could provide deeper insights into their impact across vestibular recovery stages.

Glial reactivity occurs in both the vestibular nerve and the VN in our model, suggesting a reactive inflammation [10] that could initiate the generation of new oligodendrocytes [112]. Our cUL model induced an increase in Olig2 + cells only at D3, returning to control values at D30, in line with previous findings following cortical brain injury [113]. After brain damage, increased Olig2 + cells numbers at the lesion site can result from Olig2 transcriptional up-regulation, cell proliferation or cell migration [113]. In our cUL model, we did not detect any cell proliferation in the VN, whereas others brain injury models [113] or vestibular models did [36]. This suggests that Olig2 + cell increase in our model may result from transcriptional regulation or migration. As oligodendrocytes are essential for preserving myelin integrity, their dysfunction may worsen signal transmission and delay recovery [114]. Unilateral vestibular lesions lead to decreased activity within the VN. Sensory substitution through visual and proprioceptive inputs—known to modulate VN activity [115,116]— plays a crucial role in compensation [18]. Increased activity in these sensory pathways could be reinforced by adaptive myelination [117] and could facilitate more efficient signal transmission to restore or maintain VN activity. Therefore, oligodendrocytes might reinforce substitutional input that would be beneficial to vestibular compensation. Another element to consider is that Olig2 marks oligodendrocyte progenitor cells (OPC), as well as mature oligodendrocytes [118] and that this transcription factor is plastic [119,120]. Thus, our transient increase in Olig2 + could also come from OPC migration or from OPC differentiation into oligodendrocyte. Also, an Olig2+ increase has been reported in response to stress tissue where it can be involved in oligodendrocyte lineage stabilization and cell survival [121,122]. Thus, our cUL model could induce a transient stress tissue that disappears once most of vestibular deficits are compensated.

Electrophysiological imbalance between both VN in the acute phase of vestibular syndrome may place the central nervous system in a state of functional disruption, thereby triggering rapid compensatory mechanisms involving widespread glial reactivity to restore homeostasis. In the compensated phase, however, only a strong astroglial response would persist and be sufficient to stabilize neuronal networks and maintain this electrophysiological balance that underlies vestibular compensation.

Chemical unilateral labyrinthectomy alters postural parameters and anxiety-like behavior

cUL leads to damage of vestibular hair cells, and their loss correlates with impaired vestibular function [66,123]. Consistent with our findings, other studies have reported vestibular dysfunction (such as tail hanging and landing reflexes and head tilt) following p-arsanilic acid exposure [21]. In our study, we defined three phases of the vestibular syndrome: acute, post-acute, and then compensated. Among the various qualitative assessment described in the literature, this vestibular syndrome kinetics induced by p-arsanilic acid remains similar, with symptoms peaking a few days after ototoxic exposure and plateauing one month after this ototoxic exposure [18–20,22]. Our model contrasts sharply with the peak observed at D1 in models involving vestibular nerve section [24,33,36,48,49]. As discussed previously, this difference in kinetics may be due to the sudden and complete disruption caused by surgical section [7], whereas our model induces a more gradual alteration of the vestibular sensory organs. This gradual degeneration in vestibular hair cells may account for the delayed onset of vestibular symptoms in our model, which peak a few days after the ototoxic exposure and persist for up to a month. The long-lasting nature of these deficits likely reflects permanent hair cell loss on the side exposed to the ototoxic compound. Despite adaptative plasticity and the normalization of contralateral calyces’ number, such compensatory mechanisms may not be sufficient to fully compensate for vestibular deficits induced by cUL.

Static postural deficits following a vestibular lesion are thought to result from the electrophysiological imbalance between the VN, whereas dynamic vestibular compensation mainly relies on sensory substitution mechanisms [124]. In our model, the support surface area was significantly enlarged from the first post-cUL day and remained above preoperative values at D30, as previously reported [33,51]. Interestingly, while the peak in support surface area occurred at D3 in nerve section models, it was delayed until D7 in our cUL model, which aligns with a more gradual degradation of vestibular sensory inputs. Postural instability is observed immediately after the ototoxic exposure, as evidenced by a sustained increase in body sway over time. In response, the animal adopts an adaptive postural strategy characterized by a shift in weight distribution toward the side ipsilateral to ototoxic exposure—however, this shift does not fully resolve the instability. This postural bias appears as early as D4, contrasting with nerve section models where it typically emerges later, between D7 and D10 [33,36,37,48], suggesting that such postural strategy is conserved across various vestibular deafferentation models. As more weight is applied to the ipsilateral paws relative to ototoxic exposure, the center of gravity is displaced toward the deafferented side, potentially accounting for the increased use of the ipsilateral forepaw as a stabilizing anchor. However, the weight distribution shift strategy appears maladaptive, as it is abolished by pharmacological treatments [36,49] or rehabilitation protocols [48] that promote vestibular compensation.

Circling behavior, a rapid rotational behavior linked to cerebral asymmetry, is frequently observed in rodent models of neurological disorders such as Parkinson’s disease, anxiety and vestibular impairments [5,33,125–127]. A similar behavior has also been reported in humans [128,129]. In our study, circling did not return to preoperative values at D30, despite the restoration of electrophysiological balance between both VN. This persistence may reflect the onset of hyperactivity induced by vestibular deafferentation, typically emerging after the acute phase (from D9 onward). As for the nerve section model [130], we also observe a hyperactivity phase at the same delay during our LDBT assessment with frenetic back and forth between both compartments. Several hypotheses may account for the hyperactivity observed during the compensated phase. This behavior could be a compensatory mechanism to compensate for the lack of exploration in animals with vestibular damage during the acute phase, which results in suboptimal arousal levels and reduced sensory feedback [131]. Increasing walking speed may also serve as a compensatory strategy to stabilize balance, as reported in both rodents and humans [130,132]. In our model, animal display immobility in LDBT during the acute phase, consistent with previous findings using Ethovision® [18,130]. The absence of early hyperactivity and predominantly static posture during the acute phase of the vestibular syndrome may represent an avoidance strategy to fall, similar to that observed in humans with vestibular disorders [132]. Additionally, peripheral vestibular loss may not be the only trigger for locomotor hyperactivity which may emerge through long lasting complex changes in the basal ganglion and motor associated regions that takes time [60].

Regarding anxiety-like behavior, animals consistently spent less time in the light compartment across all phases of vestibular compensation. During the acute stage, this may reflect their predominantly static posture where the animals favored a quadruped posture. In contrast, during the compensated phase, despite exhibiting hyperactivity behavior and increased velocity, animal still avoided the light compartment, suggesting persistent anxiety-like behavior, as previously reported [39,133].

Clinical relevance and limitations

Signs of inflammation of the vestibular nerve, but not of Scarpa’s ganglion, have been described in patients with acute unilateral vestibulopathy [12,56], and are replicated in our cUL model [10,22]. Acute inflammation may initially be beneficial by promoting tissue repair through tightly regulated cytokine signaling [134–136]. In contrast, excessive or prolonged glial activation can lead to chronic neuroinflammation associated with persistent tissue alterations and degenerative processes [134,135]. It is therefore plausible that chronic inflammation may be present in patients who fail to recover or compensate, raising questions about whether glial reactivity is tightly regulated and whether it could be therapeutically targeted, potentially in a patient-specific manner. In acute unilateral vestibulopathy, elevated levels of tumor necrosis factor α (TNF-α) and C-reactive protein (CRP) have also been reported [137,138]. However, these findings are not consistently observed across studies, leaving open questions about inflammation as a core feature of this disorder. This issue is clinically relevant, particularly when considering the potential use of anti-inflammatory treatments in acute unilateral vestibular disorders [56]. In this context, TSPO ligand imaging represents a promising approach to investigate glial reactivity in the vestibular nerve and nucleus of patients following acute unilateral vestibulopathy as an individualized biomarker, as has been established for patients with vestibular schwannoma [139,140].

We hypothesized that peripheral plasticity after cUL requires central detection of the lesion and may involve the efferent vestibular system (EVS) as a key modulator of bilateral peripheral vestibular plasticity. Mice with genetic alterations in cholinergic EVS signaling exhibit baseline posturo-locomotor deficits [141,142]. Moreover, EVS has been implicated in vestibular compensation of the vestibulo-ocular reflex following unilateral vestibular lesion [143–145]. Although the EVS remains poorly characterized in humans, recent reviews suggest a broader functional role than previously appreciated [146].

Finally, it should be noted that permanent unilateral hearing loss induced by cUL is inconsistent with the clinical definition of acute unilateral vestibulopathy in humans [147], which excludes hearing loss. This difference should be considered when extrapolating our findings the human condition. In addition, SHAM surgery induced moderate ipsilateral hearing loss due to tympanic membrane perforation. While this may contribute to some behavioral measures, it is unlikely to account for the pronounced vestibular deficits specifically observed after cUL.