Vestibular Hair Cells Require CAMSAP3, a Microtubule Minus-End Regulator, for Formation of Normal Kinocilia

Abstract

Kinocilia are exceptionally long primary sensory cilia located on vestibular hair cells, which are essential for transmitting key signals that contribute to mammalian balance and overall vestibular system function. Kinocilia have a "9+2" microtubule (MT) configuration with nine doublet MTs surrounding two central singlet MTs. This is uncommon as most mammalian primary sensory cilia have a "9+0" configuration, in which the central MT pair is absent. It has yet to be determined what the function of the central MT pair is in kinocilia. Calmodulin-regulated spectrin-associated protein 3 (CAMSAP3) regulates the minus end of MTs and is essential for forming the central MT pair in motile cilia, which have the "9+2" configuration. To explore the role of the central MT pair in kinocilia, we created a conditional knockout model (cKO), Camsap3-cKO, which intended to eliminate CAMSAP3 in limited organs including the inner ear, olfactory bulb, and kidneys. Immunofluorescent staining of vestibular organs demonstrated that CAMSAP3 proteins were significantly reduced in Camsap3-cKO mice and that aged Camsap3-cKO mice had significantly shorter kinocilia than their wildtype littermates. Transmission electron microscopy showed that aged Camsap3-cKO mice were in fact missing that the central MT pair in kinocilia more often than their wildtype counterparts. In the examination of behavior, wildtype and Camsap3-cKO mice performed equally well on a swim assessment, right-reflex test, and evaluation of balance on a rotarod. However, Camsap3-cKO mice showed slightly altered gaits including reduced maximal rate of change of paw area and a smaller paw area in contact with the surface. Although Camsap3-cKO mice had no differences in olfaction from their wildtype counterparts, Camsap3-cKO mice did have kidney dysfunction that deteriorated their health. Thus, CAMSAP3 is important for establishing and/or maintaining the normal structure of kinocilia and kidney function but is not essential for normal olfaction. Our data supports our hypothesis that CAMSAP3 is critical for construction of the central MT pair in kinocilia, and that the central MT pair may be important for building long and stable axonemes in these kinocilia. Whether shorter kinocilia might lead to abnormal vestibular function and altered gaits in older Camsap3-cKO mice requires further investigation.

Keywords: CAMSAP3; gait; kidney dysfunction; kinocilia; vestibular function; “9+2” configuration.

FIGURE 1

Creation and validation of the Camsap3-cKO mouse. (A) A schematic diagram of the Camsap3 knockout first targeting strategy and conversion of tm1a (KO first) to tm1c (floxed), and tm1d (cKO) alleles, respectively. TgPax2^Cre/+; Camsap3^{tm 1c/tm1c} mice, known as Camsap3-cKO, are a cKO model for the vestibular organ. (B) The exon 7 of Camsap3 execution was validated by PCR genotyping. The inner ear genomic DNA of offspring derived from crossing TgPax2^Cre and Camsap3^{tm 1c/tm1c} were isolated and used for PCR (n = 9 from two separate litters). No DNA and floxed allele Camsap3^{tm 1c/tm1c} were used as the negative and positive controls. The bands were expected to be 1307 bp for WT (TgPax2^Cre+/+; Camsap3^{tm 1c/tm1c}) and 199 bp for Camsap3-cKO mice. (C–G) CAMSAP3 protein expression in the vestibular system as examined by immunofluorescence. (C,D) Representative immunofluorescent images of a crista from WT (C) and Camsap3-cKO (D) mouse at P75 were shown. The dashed line outlines the edge of the crista of a Camsap3-cKO mouse. Scale Bars = 50 μm. (E) Vestibular cells from Camsap3-cKO mice had significantly less CAMSAP3 staining than their WT counterparts. The bars represent mean ± SD. *Statistically significant difference (p = 0.03). WT and Camsap3-cKO mice littermate (P75) replicates (n) were as indicated. (F,G) Representative confocal maximum projection of z-stack images taken from whole-mounts utricle samples of a WT (F) and a Camsap3-cKO mouse (G). More CAMSAP3 staining dots were found on the apical surface of WT utricle compared to those on Camsap3-cKO mouse. Scale Bars = 10 μm. Antibodies: anti-Camsap3 (Green), phalloidin (Red).

FIGURE 2

Kinocilia on utricles hair cells from Camsap3-cKO are shorter than their WT littermates. (A) A representative 3D-reconstruction from z-stacks images of bundles of stereocilia (red) and kinocilia (green) located on the apical surface of hair cells from an utricle of a WT mouse (P386). Antibodies: anti-acetylated tubulin (Green), phalloidin (Red). The dashed line outlines the edge of actin-based stereocilia bundles. Scale Bar = 10 μm. (B) A schematic diagram showing the region over which Z-stack images of utricle hair cells were collected. (C) Quantification of kinocilia length collected from male WT (n = 3) and male Camsap3-cKO littermates (n = 3) with ages ranging from 10- to 13-month. Each dot represented one kinocilium. Bars represent mean ± SD. Kinocilia on utricles hair cells from Camsap3-cKO were statistically significant shorter than WT (p < 0.0001). *Statistically significant difference.

FIGURE 3

CAMSAP3 contributes to the formation of central MT pairs in axonemes of kinocilia on utricle hair cells. (A) The overlaps between CAMSAP3 (green) and basal bodies (red) were displayed by a group of consecutive Z-stack kinocilia images taken from a whole mount utricle sample of WT (P42). Z-stack images were captured using the optical section (0.5 μm) starting from the apical surface of utricle hair cells toward its nucleus. Antibodies: anti-CAMSAP3 (Green), anti-γ-tubulin (red), phalloidin (white). Scale Bars = 5 μm. Two basal bodies of kinocilia were indicated by arrows. (B) A schematic diagram illustrates the region over which Z-stack images of kinocilia immunostaining A, and TEM images C-D were collected. (C,D) Representative TEM images show transverse sections of kinocilia on utricle hair cells from 8-month-old WT with a “9+2” configuration (black arrow) (C), and a “9+0” configuration (black arrow) for their Camsap3-cKO littermates (D). Scale Bars = 1 μm. (E) The distribution of MT arrangements for both WT and Camsap3-cKO mice was compared. Distribution of the MT configurations between WT and KO was significantly different as analyzed by a Kolmogorov–Smirnov test (p = 0.0003).

FIGURE 4

Weight differences between WT and Camsap3-cKO mice. (A) Male Camsap3-cKO mice weigh less than their WT counterparts after the age of 8-month, while no body weight difference was observed in mice younger than 4-month old. Mice were weighed prior to euthanasia at ages P36–P388. ns, not statistically significant difference. (B) Female Camsap3-cKO mice aged 8–13 months weigh less than their WT counterparts. For both male and female mice, the weights of old WT and Camsap3-cKO mice were significantly different. p-values for each age group of mice were as indicated. *Statistically significant difference.

FIGURE 5

Gait differences between WT and Camsap3-cKO mice. Two litters of mice, ages P130–P133, were tested using the DigiGait Imaging System. Multiple t-tests revealed that Paw Area (A,B) and Max dA/dt (C,D) were significantly different between WT and Camsap3-cKO mice (p-values for each speed were as indicated). Bars represent mean ± SD. (E) Unpaired t-tests showed that there was no body length difference between male WT and Camsap3-cKO mice (ns, p = 0.1572). (F) Unpaired t-tests showed that there was also no body width difference between male WT and Camsap3-cKO mice (ns, p = 0.0539). N = 3 for both WT and Camsap3-cKO mice. *Statistically significant difference.

FIGURE 6

Renal abnormalities in Camsap3-cKO mice. (A,B) The kidney-to-body weight ratio in male (A) and female (B). Unpaired t-tests revealed that Camsap3-cKO mice (P22-P339) showed a statistically significant increase in their kidney-to-body weight ratio relative to their WT counterparts. p-values for each age group of mice were as indicated. Bars represent mean ± SD. (C,D) Representative kidney images taken from male, 1-year-old WT (C) and Camsap3-cKO littermate (D). Kidneys from the Camsap3-cKO mouse were hypertrophic and discolored. *Statistically significant difference.

FIGURE 7

WT and Camsap3-cKO mice show no differences in olfaction. (A) No differences in the reaction time of the sense of smell between WT and Camsap3-cKO mice. Bars represent mean ± SD. Unpaired t-tests revealed no significant (ns) differences between WT and Camsap3-cKO mice in male or female mice. (B,C) CAMSAP3 had little expression in olfactory bulbs regardless of genotype, WT (B) or Camsap3-cKO (C). The representative images were taken from female WT and their Camsap3-cKO littermates at P180. (B’,C’) were the cells shown in (B,C) stained with anti-Camsap3 (green). (D–G) CAMSAP3 was expressed in olfactory sensory neurons (OSN) in both WT and Camsap3-cKO mice. Olfactory mucosa in the nasal cavity from WT (D,F) and Camsap3-cKO (E,G) mice were stained with anti-Camsap3 (green), anti-acetylated tubulin (red, olfactory cilia, and nerve fibers), and Hoechst 33342 (blue, nuclei). (F,G) were images with higher magnification showing olfactory cilia and dendritic knobs layers on OSNs. (D’,E’,F”,G”) Showed the anti-CAMSAP3 (green) channel in (D–G). (F’,G’) Showed anti-acetylated tubulin (red) channel in (F,G). “[” Marks the OSN layer in the olfactory epithelium. Scale Bars = 100 μm (B,C), 50 μm (D,E), 10 μm (F,G).