Inner ear dysfunction in caspase-3 deficient mice

Background Caspase-3 is one of the most downstream enzymes activated in the apoptotic pathway. In caspase-3 deficient mice, loss of cochlear hair cells and spiral ganglion cells coincide closely with hearing loss. In contrast with the auditory system, details of the vestibular phenotype have not been characterized. Here we report the vestibular phenotype and inner ear anatomy in the caspase-3 deficient (Casp3-/-) mouse strain. Results Average ABR thresholds of Casp3-/- mice were significantly elevated (P < 0.05) compared to Casp3+/- mice and Casp3+/+ mice at 3 months of age. In DPOAE testing, distortion product 2F1-F2 was significantly decreased (P < 0.05) in Casp3-/- mice, whereas Casp3+/- and Casp3+/+ mice showed normal and comparable values to each other. Casp3-/- mice were hyperactive and exhibited circling behavior when excited. In lateral canal VOR testing, Casp3-/- mice had minimal response to any of the stimuli tested, whereas Casp3+/- mice had an intermediate response compared to Casp3+/+ mice. Inner ear anatomical and histological analysis revealed gross hypomorphism of the vestibular organs, in which the main site was the anterior semicircular canal. Hair cell numbers in the anterior- and lateral crista, and utricle were significantly smaller in Casp3-/- mice whereas the Casp3+/- and Casp3+/+ mice had normal hair cell numbers. Conclusions These results indicate that caspase-3 is essential for correct functioning of the cochlea as well as normal development and function of the vestibule.


Background
Caspase-3 is one of the most commonly shared downstream executioners in different apoptotic pathways. In the inner ear, apoptosis and molecules involved in the apoptosis pathways play an important role during development [1,2] and in response to stress in adulthood. Caspase-3 activation is induced in response to ototoxic stress such as aminoglycoside antibiotics, cisplatin or noise, which in turn leads to hair cell death [3,4].
Mice deficient of molecules in the apoptotic pathway are useful tools to investigate their direct molecular role (s) in the inner ear. Caspase-3 deficient mice, although reported to have neuronal defects and a short life span, have no other obvious developmental abnormalities [5]. Back-crossing of the Casp3 mutant mice to the C57BL/ 6 strain, resulted in a longer lifespan without significant gross anatomical abnormalities [6][7][8]. These mice were found to have a smaller body size and to exhibit decreased hearing, hyperactivity, and circling behavior, suggestive of inner ear dysfunction.
Detailed characterization of the auditory phenotypes of two strains of caspase-3 deficient mice with a targeted deletion of exon 5-6, which encodes the QACRG pentapeptide motif in the catalytic domain of caspase-3, have been reported [7,8]. In these mice, although the development of the organ of Corti seems normal, degeneration of the auditory hair cells and spiral ganglion cells [7], as well as hyperplasia of supporting cells (border cells) were observed [8] coinciding with severe progressive hearing loss. In addition, Melody, an ENU mutant with a point mutation within the catalytic domain of Casp3, has been shown to have similar auditory phenotype. These mice also display severe hearing loss as well as loss of auditory hair cells and spiral ganglion cells [9].
In contrast with the detailed characterization of the auditory phenotype in these Casp3 mutant mice, details of the vestibular phenotype have not been reported.
Here we report the characterization of the vestibular phenotype in the caspase-3 deficient mouse.

Vestibular dysfunction in caspase-3 deficient mice
Most Casp3 -/mice were hyperactive and exhibited circling behavior with a tendency towards a unilateral directional rotation when excited. We observed that about 70% of the mice circle counter-clockwise (left direction), while 25% circle clockwise (right direction). We tested the horizontal vestibulo-ocular reflex (VOR) in the dark using near-infrared video-oculography to assess vestibular function in different genotypes. The frequency response of the VOR in darkness for each group was characterized during horizontal rotation at 5 frequencies ranging from 0.05Hz to 0.8Hz, with peak velocity of 60 degrees/sec. Little response at any frequency or velocity was recorded for Casp3 -/mice (n = 10) (Figure 2, 3). Both heterozygous Casp3 +/-(n = 13) and wild type Casp3 +/+ (n = 14) mice exhibited a high-pass filtered VOR response, with higher gains and lower phase errors with increasing stimulus frequency ( Figure 3A, B). The VOR at 0.2 Hz was also linear over the range of peak velocities from 60 to120 deg/sec for both groups ( Figure 3C). The gain of the Casp3 +/mice was intermediate between Casp3 -/and Casp3 +/+ mice (P < 0.05).

Inner ear dysmorphism in caspase-3 deficient mice
Anatomical analysis of the inner ear of Casp3 -/mice revealed gross hypomorphology in the vestibule, mainly in the anterior semicircular canal (Figure 4). Various Averages of distortion product 2F1-F2 at F2 frequencies of 10 kHz, 11.3 kHz, and 12.3 kHz were compared at 3 months of age. Casp3 -/-(n = 5) mice had a significantly smaller distortion product (*, P < 0.05) compared to Casp3 +/-(n = 9) and Casp3 +/+ (n = 5) mice. degrees of malformations in the anterior semicircular canal were observed in 14 out of 15 Casp3 -/mice tested, whereas all of the Casp3 +/mice (n = 20) or Casp3 +/+ (n = 14) had normal gross semicircular canal anatomy. The most common malformation was a decreased arc size of the anterior semicircular canal (n = 10). Other severe malformations included truncation or aplasia of the anterior semicircular canal. Those with severe anterior semicircular canal malformations were often accompanied by a hypomorphic lateral semicircular canal (n = 3), but to a much smaller frequency. However, the posterior semicircular canal was always preserved. Most malformations were observed unilaterally in the left ear (n = 6) or in the right ear (n = 3), but also in bilateral ears (n = 5). Gross anatomy of the cochlea appeared normal in all genotypes ( Figure 4). Histological studies of sensory epithelia were consistent with gross anatomy findings. The crista of the anterior-and lateral semicircular canals was hypomorphic, whereas that of the posterior semicircular canal developed normally ( Figure 5). Hair cell numbers were significantly smaller (P < 0.05) in the anterior-and lateral crista, and in the utricle, in Casp3 -/mice (n = 7), whereas Casp3 +/mice (n = 5) and Casp3 +/+ mice (n = 3) had normal hair cell numbers in all vestibular sensory epithelia ( Figure 5F). While the size of the hair cells looked normal, the decreased number of hair cells seemed to be the main reason for the hypomorphic anterior crista ampullaris in Casp3 -/mice ( Figure 5E). In some Casp3 -/mice, the anterior crista was absent, or fused with the lateral crista ( Figure 5C). All Casp3 +/mice and Casp3 +/+ mice observed to date had normal morphology and hair cell numbers in the vestibular sensory epithelia. Although most Casp3 -/mice seemed to circle toward the more severely affected side of the ear, there was no significant difference between the hair cell number in the left ear and right ear (Student's t test, P > 0.05).

Caspase-3 is essential for the maintenance of normal hearing
When different regions are targeted to delete a gene of interest, the knockout mice may exhibit different phenotypes depending on the background strain, even though the same gene is deleted [10]. The auditory function test results from ABR and DPOAE of our Casp3 -/mice were consistent with previously reported profound hearing loss in two different strains of Casp3 knockout mice [5,7]. The similar phenotype in the different mutant strains of Casp3 (Table 1)  caspase-3 in the maintenance of normal hearing after the developmental stages of the cochlea.

Caspase-3 is important for development of the vestibule
Most mutant mice with inner ear phenotypes have extensive developmental defects in the vestibular organs [11]. The lateral semicircular canal is one of the most vulnerable, and morphologically affected [12]. On the other hand, those with either isolated anterior-or posterior semicircular canal defects are rare (Table 2). Apaf1 knockout mice and Casp9 knockout mice are a few of such mutants with specific anterior semicircular canal defect [13]. Interestingly, both apaf1 and caspase-9 are molecules involved in apoptosis. Apaf1 together with caspase-9, form a multiprotein complex called apoptosomes, which in turn activates downstream caspase-3, in the apoptotic pathway [14,15]. In these knockout mice, apoptosis is greatly reduced during development of the inner ear [13]. One possible mechanism leading to the phenotype is speculated to be the reduction of apoptosis, which in turn decrease the passive release of functional factors from the dying cells into the local environment [13]. Only a few genes are known to be exclusively expressed in a specific crista or canal during development [16]. There is no gene reported to be expressed specifically in the anterior crista or canal. Although many genes are ubiquitously expressed, differential expression of genes are required for the formation of each of the three cristae and canals [16]. This suggests an important role for this particular apoptotic cascade involving apaf1-caspase-9-caspase-3 in the development of the anterior semicircular canal. Apoptosis plays an important role throughout the development of the inner ear [1], including the stages during the innervation of the vestibular epithelia from the sensory ganglion and the differentiation of the otic epithelia into the sensory epithelia [2]. In Casp3 -/mice, in contrast to the normally developed cochlea, in the vestibule, the anterior crista, lateral crista, and utricle were hypomorphic at birth. The gross innervation to these three vestibular organs was reduced in our preliminary observation of the nerve fibers and vestibular ganglion cells. The anterior crista, lateral crista, and utricle are innervated by the superior vestibular nerve [17]. This may suggest that caspase-3 is one of the major apoptotic molecules involved in the development of the superior vestibular nerve, which innervates all three vestibular organs.
Either as a direct-or indirect consequence, caspase-3 seems to have an important role in the development of the vestibule, especially in the anterior-and lateral semicircular canals.

Vestibular dysfunction in caspase-3 mutant mice
The horizontal VOR paradigm tested the vestibular ocular pathway for the lateral semicircular canal. The VOR results of the Casp3 +/+ mice were qualitatively similar to other studies in mice using both search coil [18] and videographic methods [19]. In our Casp3 -/mice, even those with relatively normal lateral semicircular canal morphology, had little response to any stimuli tested. Vidal et al. [20] showed that the mutant mice with deficient VOR function also had profound deficits in locomotion. It is not yet clear the extent to which peripheral versus central mechanisms contributes to the "circling behavior." The unidirectional circling behavior towards the more severely anomalous vestibule suggests that the vestibular system is functional to some degree postnatally, corresponding to the extent of hypomorphism of the vestibule. Further functional and histological studies on the inner ear and central vestibular pathways would be needed to clarify these questions.
Although Casp3 +/did not exhibit any abnormal behavior or gross anatomical abnormalities, VOR analysis revealed an intermediate phenotype. Stimulation with rapid frequencies produced significant differences between Casp3 +/and Casp3 +/+ mice, suggesting that adequate vestibular development has not occurred in Casp3 +/mice.
In Casp3 +/mice, we observed slightly increased hair cell numbers in most vestibular organs compared to Casp3 +/+ mice ( Figure 5F). In mammals, regeneration has been observed in vestibular hair cells after damage from ototoxic agents [21][22][23]. Perhaps in the Casp3 mutants, caspase-3 insufficiency causes overall decrease in cell death at later stages of development or even after birth. One possible mechanism of mild vestibular dysfunction may be that perhaps in Casp3 +/mice there is an imbalance in cell population and corresponding neurons which caused improper functional development.

Conclusions
The results in this study, together with previous inner ear studies on caspase-3 deficient mice, indicate that (1) caspase-3 is essential for normal function of the auditory and vestibular system, and (2) caspase-3 is important for the morphogenesis of the anterior semicircular canal. These mice can be useful tools for further studies, such as characterizing morphogenesis of the anterior semicircular canal, or as a control animal for developing vestibular function tests in mice.

Mice
Caspase-3 deficient mice were a gift from Minna Woo (University of Toronto). The generation of the CPP32 ex3-/-( = Casp3 -/-) are described [6]. The Casp3 gene was disrupted by targeted deletion of exon 3, which leads to termination codons in all three reading frames [6]. The mice were backcrossed extensively with the C57BL/6 strain. The homozygous mice were born at a lower than expected Mendelian frequency (9%), and were smaller than their littermates initially [6], but caught up in size around 2 months of age and were reproductive. The mice were maintained and bred at the Animal Resource Center at University of Texas Medical Branch. All possible steps were taken to avoid animals from suffering as well as keeping the number of animals used in the studies to a minimum. All of the experiments and procedures were approved by the Institutional Animal Care and Use Committee at University of Texas Medical Branch.

Auditory function testing
For auditory function testing, the mice were anesthetized by intraperitoneal injection of pentobarbital (60 mg/kg body mass). The mice were kept on a heating pad for body temperature control during all auditory tests. The auditory tests were performed in a sound proof environment. Wild type (Casp3 +/+ ), heterozygous (Casp3 +/-) and homozygous (Casp3 -/-) mice at 3 months of age were tested, and the results were compared among different genotypes. For ABR, subcutaneous needle electrodes were placed adjacent to the left and right pinna and at the occipital region. Sound stimuli were delivered to the ear by an ear probe. ABR recordings were performed using Nicolet CA2000 system and software (Cardinal Health, Madison, WI, U.S.A.). Averages of 500 responses to click stimuli were recorded for descending 5 dB stimulus steps to determine the threshold. Averages of the betterhearing ear were compared among the different genotypes.

Vestibular function testing
For vestibulo-ocular reflex (VOR) testing, we used a custom-made rodent centrifuge [25], modified for use in mice. Details are described [25], except for modifications as follows. After brief isoflurane inhalation anesthesia, the mouse body is restrained with a plastic cone and then secured onto a custom-made bed. The anterior portion of the head including the pinna is placed outside of the narrow end of the cone. Then the nose and maxilla is secured with a bite block, with the head restrained with a custom-made adjustable head-mount. The nose is angled in a 35°nose down position, resulting in an earth horizontal plane for stimulation of the lateral semicircular canal. Forearms are restrained inside the plastic cone, and hind legs are extended, restrained, and secured with a custom-made leg restraint. Bite block and bed are attached to a mini-platform that is designed to dock into the centrifuge to assure stability of location. After securing the mouse into the centrifuge, the mouse was kept quiet for 10-15 minutes to assure full awakening and to stabilize before the test is started. Mice were tested at 3 months of age. Eyes were treated with 1% pilocarpine to restrict pupils. Horizontal eye position was derived from tracking the pupil center using a least squares fit to a circular disk model [26]. Calibration scale factors were geometrically derived based on the ocular globe image size, and verified by rotating a camera about the midpoint of the globe. Following differentiation and fast phase removal, nonlinear least squares sinusoidal curve fits to the remaining slow components of horizontal eye velocity were used to determine the gain and phase of the horizontal canal VOR over at least three successive cycles. The frequency response of the VOR in darkness was characterized during horizontal rotation at five frequencies ranging from 0.05 Hz to 0.8 Hz, with peak velocity of 60 deg/sec. The linearity of the VOR was tested at 0.2 Hz using peak velocities ranging from 60 to 120 deg/sec.

Anatomical studies of the inner ear
The mice were euthanized with CO 2 inhalation followed by decapitation following the IACUC protocol. The inner ears were dissected from the temporal bones, and were fixed with 4% paraformaldehyde in phosphate buffered saline (PBS) for more than 2 hours. The surrounding soft tissue and mastoid air cells were carefully cleaned using fine forceps. Inner ears from mice were collected at different ages up to 6 months of age.
Whole mount preparation of the inner ear Mice aged 6 months or younger were used. The mice were sacrificed with CO 2 inhalation followed by decapitation. The temporal bone was immediately dissected and perfused with 4% paraformaldehyde in PBS for 1 hour. Then the organ of Corti and vestibular sensory epithelia were carefully dissected. The specimen were washed with PBS for 5 minutes, permeabilized with 0.5% Triton X-100 for 30 minutes, and again washed with PBS for 5 minutes x3. Then, under dark conditions, the specimen was incubated with Alexa fluor 568-conjugated Phalloidin diluted in PBS for 60 minutes, and then was incubated with Hoechst33342 diluted in PBS for 15 minutes to stain the hair cell stereocilia and nuclei, respectively. After washing with PBS for 5 minutes x3, the specimen was mounted with Slowfade Gold antifade mounting solution (Invitrogen) and sealed. The inner ear sensory epithelia were observed under fluorescent microscopy (Nikon TE2000) at 100x, 150x, 400x, 1000x, and 1500x magnifications. Images were captured using Nikon NIS element software (Nikon).
Hair cell count: Cells with stereocilia labelled with Alexa fluor 568 phalloidin at the apical surface were counted. A combination of 150x and 400x magnifications were used to extract individual cells as accurate as possible for counting.

Statistical analysis
Results from the auditory and vestibular testing were compared among Casp3 +/+ , Casp3 +/and Casp3 -/mice. Statistical difference (P < 0.05) was assessed by ANOVA, or Student's t test using the StatistiXL version 1.6 (available from http://www.statistixl.com/) with the Windows™ version of Microsoft Excel™.