Implication of Vestibular Hair Cell Loss of Planar Polarity for the Canal and Otolith-Dependent Vestibulo-Ocular Reflexes in Celsr1-/- Mice

Abstract

Introduction: Vestibular sensory hair cells are precisely orientated according to planar cell polarity (PCP) and are key to enable mechanic-electrical transduction and normal vestibular function. PCP is found on different scales in the vestibular organs, ranging from correct hair bundle orientation, coordination of hair cell orientation with neighboring hair cells, and orientation around the striola in otolithic organs. Celsr1 is a PCP protein and a Celsr1 KO mouse model showed hair cell disorganization in all vestibular organs, especially in the canalar ampullae. The objective of this work was to assess to what extent the different vestibulo-ocular reflexes were impaired in Celsr1 KO mice. Methods: Vestibular function was analyzed using non-invasive video-oculography. Semicircular canal function was assessed during sinusoidal rotation and during angular velocity steps. Otolithic function (mainly utricular) was assessed during off-vertical axis rotation (OVAR) and during static and dynamic head tilts. Results: The vestibulo-ocular reflex of 10 Celsr1 KO and 10 control littermates was analyzed. All KO mice presented with spontaneous nystagmus or gaze instability in dark. Canalar function was reduced almost by half in KO mice. Compared to control mice, KO mice had reduced angular VOR gain in all tested frequencies (0.2-1.5 Hz), and abnormal phase at 0.2 and 0.5 Hz. Concerning horizontal steps, KO mice had reduced responses. Otolithic function was reduced by about a third in KO mice. Static ocular-counter roll gain and OVAR bias were both significantly reduced. These results demonstrate that canal- and otolith-dependent vestibulo-ocular reflexes are impaired in KO mice. Conclusion: The major ampullar disorganization led to an important reduction but not to a complete loss of angular coding capacities. Mildly disorganized otolithic hair cells were associated with a significant loss of otolith-dependent function. These results suggest that the highly organized polarization of otolithic hair cells is a critical factor for the accurate encoding of the head movement and that the loss of a small fraction of the otolithic hair cells in pathological conditions is likely to have major functional consequences. Altogether, these results shed light on how partial loss of vestibular information encoding, as often encountered in pathological situations, translates into functional deficits.

Keywords: CELSR1; hair cell; mouse model; planar cell polarity (PCP); vestibular system; vestibulo ocular reflex.

FIGURE 1

Hair cell orientation of semicircular canal ampullae and otolithic maculas in wild-type and celsr1^{− /−} mice. In semicircular canals, the ampullae (A1) are stimulated following a rotation, whereas (B1) the utricule and saccule are stimulated following translational movements. The organization of hair cells in the ampulla (A2) are normally all orientated in the same direction, whereas in celsr1^{− /−} mice, 80% of hair cells are misorientated. The organization in the macula (B2) is more complex with hair cells orientated toward the striola (in the utricule shown here), and moderate misorientation in celsr1^{− /−} mice. Hair cell orientation of wild type (WT) and celsr1^{− /−} are schematically represented based on results from Duncan et al. (2017). Transmission electronic microscope of the saccular macula of a KO mouse. The × 600 (C1) image shows an overall maintained saccular structure with otoconia and sensory hair cells beneath; (C2) × 5,000 shows mostly stereociliary bundles and (C3) × 5,500 shows a clear misalignment of two neighboring hair cells, confirming the previous study (Duncan et al., 2017).

FIGURE 2

Semicircular canal function. The horizontal semicircular canal (SCC) was tested. Horizontal sinusoidal rotation (A) tested for horizontal aVOR. Traces in (A1) show head position (corresponding to table position) and horizontal eye movement. KO mice show clear reduction in the amplitude of the eye movement. aVOR gain (A2) for maximum 30°/s velocity showed reduced gain for all frequencies of stimulation. aVOR phase (A3) for maximum 30°/s velocity showed increased phase for all frequencies of stimulation except 1.5 Hz. Horizontal steps (B1) at 50°/s were performed. Traces show head position and horizontal eye position, during per-rotatory nystagmus at the start of a CCW stimulation. In this example trace, only two fast phases can be identified at velocity changes for the KO mouse compared to seven for the WT. Time constant (B2, left) and peak gain (B2, right) are both decreased for Celsr1 KO mice, for hsteps at 50°/s, for per- and post-rotatory nystagmus during CW and CCW steps. aVOR angular vestibulo-ocular reflex, CW Clockwise, CCW Counterclockwise, hsteps horizontal steps, WT wild type. *p < 0.05; **p < 0.01; ***p < 0.001.

FIGURE 3

Otolithic system function. Static roll head tilts (A1) were performed to determine OCR testing utricular function. Traces show head position (corresponding to table position) and vertical eye movement. KO mouse had preserved function with about 1/3 poorer gain. Vertical eye position in degrees was calculated for OCR according to roll head tilt degree. The static left eye pupil position was measured; tilts with positive degrees were toward the left side and those with negative degrees were toward the right side. OCR gain (A2) was significantly reduced in KO mice. The dynamic (gradient fill) head tilt gain was compared to the static (solid fill) OCR gain (A3). In KO mice, the amplitude was not significantly changed between static and dynamic stimulations, whereas it was in WT mice. The Maculo-ocular reflex was studied with the off-vertical axis rotations (B1) at 50°/s. Traces show head position and horizontal eye position, during CW stimulation. KO mouse in this example has poor response, seen as the absence of the MOR nystagmus. We report MOR bias (B2, left), which is significantly reduced in KO mice, and modulation (B2, right), which is similar in both groups. CW Clockwise, CCW Counterclockwise, MOR Maculo-ocular reflex, OCR Ocular counterroll, OVAR Off-vertical axis rotation, WT wild type. *p < 0.05; **p < 0.01; ***p < 0.001.

FIGURE 4

Saccade main sequence. At least 15 fast phases generated by WT (n = 10) and KO mouse (n = 10) during OVAR stimulation were analyzed, for a total 159 and 186 fast phases, respectively. The amplitude, duration, and peak velocity were quantified using a 20°/s threshold for saccade onset and offset (Beraneck and Cullen, 2007). (A) Shows the relation between amplitude and duration. For WT, linear regression parameters were: Y = 0.0003x + 0.032; R² = 0.024; F(1, 157) = 3.93, p = 0.049. For KO, linear regression parameters were: Y = 0.0008x + 0.030; R² = 0.055; F(1, 184) = 10.64, p = 0.001. Regression were not significantly different [F(1, 341) = 2.53, p = 0.113]. (B) Shows the relation between amplitude and velocity. For WT, linear regression parameters were: Y = 25.38x + 44.02; R² = 0.607; F(1, 155) = 239.3, p < 0.0001. For KO, linear regression parameters were: Y = 16.08x + 140.7; R² = 0.242; F(1, 184) = 58.59, p < 0.0001. Regression were significantly different [F(1, 339) = 12.24, p < 0.001].