Introduction
The zebrafish (
Danio rerio) is a popular vertebrate model for auditory studies that encompass heredity, development, damage, regeneration, and drug discovery [
–
]. The number of auditory literature involving zebrafish has rapidly increased over the past 20 years (Fig. 1A). In contrast, the papers involving the guinea pig, a classical auditory model, gradually declined in number (Fig. 1B). The rapid growth in the scientific use of zebrafish is attributed to its similarities to humans in terms of gene and drug responses, ease of access to its hearing structure, and other common features. Despite the advantages of the zebrafish model, rodents including guinea pigs are still widely selected as the primary auditory model slightly because of habitual preference. In the past year, 33 published auditory studies involved zebrafish, whereas the studies involving guinea pigs tallied 197.
A major factor that limits the use of zebrafish in auditory research is the deficiency in the methodology of assessing auditory function. The auditory function in mammalian vertebrates is typically measured by electrophysiological techniques, including auditory evoked potentials, microphonic potential, and patch clamp determination of cell membrane potential. These measurements are objective and precise but are difficult to conduct on zebrafish and its larva because of their miniature size. A popular method for assessing the hearing is the microscopic imaging to inspect neuromast hair cells, neuro-sensory cells that detect balance, vibration, and current flow.In vivo imaging is straightforward in zebrafish because of its easy accessibility to neuromasts that are found at stereotypic positions along their lateral line with hair cells aggregated into clumps. However, a functional assessment is still necessary because the imaging can not be repetitively applied on the same individual animal over a long period, and the morphological alternation is not necessarily parallel to the functional change. Visible changes in neuromast morphology indicate severe damage. However, these changes are insufficient to evaluate physiological changes that are induced by genetic modification or drug intervention. Fortunately, various behavioral methods were devised to meet this demand. For instance, three types of nervous reflexes, namely, acoustic startle response, vestibular-ocular reflex (VOR), and rheotaxis, are stable and reliable in evaluating the function of hair cells in zebrafish.
To provide guidance in selecting appropriate methods for auditory studies, we herein reviewed the literature on the behavioral tests of auditory assessments in zebrafish. The rationale of choosing zebrafish as an auditory model was first discussed. Secondly, to clarify how to acquire accurate results from zebrafish, we summarized the auditory function assessing methods, focusing on the three behavioral tests based on the startle response, the VOR, and the rheotaxis. Finally, we envisioned the scientific use of the behavioral tests in zebrafish, emphasizing their potential in the high-throughput screening of genes/drugs in the auditory system, as well as in the analysis of neural network.
Why use the zebrafish in auditory studies?
Zebrafish is one of the simplest vertebrates. In vertebrates, the auditory organ is highly conserved. Although invertebrates have responses to sound and vibration, these organisms lack the structures of the inner ear for the hearing sense [
]. Among vertebrates, birds lack genetic amenability. Hair cells of mammals are difficult to access because their inner ears are deeply embedded in the temporal bone. Maintaining the physiological activity of explanted hair cells for hours
in vitro is also difficult. In addition, not all mammals necessarily preserve drug susceptibilities similar to humans. For example, the mouse is resistant to aminoglycosides, a class of ototoxic antibiotic drugs for humans. This fact inconveniently limits their use in the evaluation of ototoxic and otoprotective drugs [
]. The zebrafish possesses a clear genetic background, completely sequenced genome, and excellent anatomical features favorable for auditory studies. Numerous experimental assays were developed for zebrafish, thereby producing massive data sets. These attributes make zebrafish an ideal candidate animal model in auditory studies.
High comparability with humans
The zebrafish is comparable to humans in terms of heredity, structure, development, and chemical sensibility aspects of the auditory system. The zebrafish genome displays 71.4% homology with humans [
]. About 50 hearing-related genes were discovered in zebrafish; many of these genes similarly influence the inner ear of humans and other vertebrates [
]. Both zebrafish and mammals rely on the same mechano-electric transducing structure of a hair cell system, in which most functional molecules, such as myosin and cadherin, maintain conservative isoforms [
]. Embryonic development of the inner ear is highly conserved in vertebrates. In this process, the rudimental ear begins with the epidermal placode, which rapidly differentiates into a vesicle and later splits into three orthogonal semicircular canals and two chambers [
,
]. Several patches of sensory epithelia reside in these compartments. The sensitivities to a variety of ototoxins, oto-protectants, and oto-regeneratives are comparative in the zebrafish and humans [
,
]. Therefore, the results of zebrafish experiments are highly translatable to humans.
Accessibility of hair cells
The zebrafish has two hair cell systems that are readily accessible: the inner ear and the lateral line. The inner ear is in a hyaline otocyst close to the body surface and is designated to sense acoustic stimuli in the water. The inner ear rapidly develops within a few days after fertilization [
]. The lateral line system also detects sound, but more essentially, this system allows zebrafish to determine the direction and speed of water movement around its body, thereby providing spatial awareness. The system comprises individual sensory patches, namely, neuromasts consisting of analogous hair cells with the mammalian inner ear in terms of morphology, function, and mechanism [
]. Neuromasts are located in the superficial positions with excellent permeability of various dyes, and each neuromast is arranged in a stereotypical pattern that is easy to microscopically detect. In addition, the zebrafish larvae develop outside the maternal body and are transparent, thereby facilitating non-invasive visualization and experimental manipulation.
High-throughput capability
The zebrafish has a series of properties that are suitable for high-throughput experiments. First, a pair of zebrafish can produce about 200 synchronously developing embryos per week. The embryos/larvae rapidly grow and sexually mature and spawn again in 3 months. Second, zebrafish are small in size. An adult zebrafish is smaller than 4 cm, whereas a larva is only a few millimeters long. Thus, zebrafish is easily transferred into multi-well plates for simultaneous assays. Third, zebrafish are highly social and can be maintained in high density. The space needed to breed a couple of mice will house hundreds of zebrafish. Lastly, zebrafish are easy to rear at a lower cost than rodents. A great number of homogeneous samples can be quickly and economically generated for large-scale screening because of the high fecundity, short reproductive time, small size, and low cost of this species.
Some intrinsic disadvantages of the zebrafish should also be recognized and alerted. The evolutionary distance between zebrafish and human requires extra awareness from research investigators before reaching clinically relevant conclusions. Clinically implicative results from zebrafish should be verified in animal models with higher phylogenetic positions. Nonetheless, the zebrafish is valuable for primary screening of genes and drug compounds, studying about the inner ear and hair cell physiology, and factor analysis of hearing damage, protection, and regeneration.
Approaches to assess auditory function
Considering that the zebrafish is becoming an important model for auditory studies, accurate measurements of the auditory function is critical.
Electrophysiology
Electrophysiological techniques are typically used for auditory assessment. These techniques are fast and accurate in mammals but are complicated and produce unsteady results in zebrafish. These technologies include the auditory-evoked potential (AEP), microphonic potential, and cellular potential via patch clamp. Since these three technologies respectively record the response of the auditory neural center, sensory organs, and cells, the electrodes are accordingly located on the skin near the brain stem, inner ear/lateral line, and hair cell/ganglion cell [
–
]. These technologies are invasive and sophisticated to be applied on miniature zebrafish larvae, yet zebrafish develops hearing ability since 5-day post-fertilization (dpf). This factor is a major drawback because most auditory malformations occur in larvae and only a minority of abnormal larvae will survive to adults. On the contrary, the patch clamp approach measures only one cell at a time and thus only records a handful cells at different states of damage and regeneration. As a result, the patch clamp technique does not assess the entire hearing function. Additionally, electrophysiological data are not appropriate for longitudinal comparison. For instance, AEP data obtained from zebrafish showed conflicting relationship between hearing sensitivity and age/size of subjects [
–
]. This phenomenon occurs because of the Weberian ossicles connecting the inner ear to the swim bladder, which play a facilitative role in enhancing hearing sensitivity and extending the upper frequency limit [
]. Therefore, the inflatable degree of the swim bladder interferes with the electrophysiological results. Yin and Wang established a baseline of hearing sensitivity for zebrafish from 40 dpf to 20 months post-fertilization, revealing that the hearing threshold decreases during development and increases during aging [
]. This outcome is consistent with the morphological changes of auditory hair cells but not with preceding studies. In addition, no expansion of the hearing frequency was observed along with the development of the Weberian ossicles. This finding is unclear for explicit reasons, results in conflictive conclusions, and thus restricts the uses of AEP in zebrafish.
Hair cell imaging
An alternative and indirect approach to estimate hearing in zebrafish is by checking the number and integrity of neuromast hair cells through vital imaging. Various imaging methods include direct observation without staining (e.g., dark field microscopy) [
], specific dyes for hair cells, and fluorescent labeling [
]. Fluorescent labeling highlights the cells, and the gradient fluorescence intensities correlate with the state of the hair cells [
]. Styryl dye and few fluorescence products were used to examine the severity of hair cell damage and revealed drug dose-dependent death and subsequent regeneration [
]. However, existing drawbacks of fluorescence imaging include phototoxicity and photoquenching, which limit the accuracy of assessment on hair cell function.
Behavioral methods
Although the main research approaches are genetic, immunological, and other technologies in biological medicine, behavioral methods are irreplaceable because of their two advantages. On the one hand, the goal of biological medicine is to achieve fine function of the whole body, which is not simply the accumulation of the function of molecules or cells, especially in studies about networks (endocrine, metabolic, and neural system). On the other hand, behavior is a fast and direct indicator in these studies.
Zebrafish harbors a rich repertoire of motor behavior from the early life stage, allowing the development of many behavioral measurements. These behavioral methods are economic and reliable for functional assessments in zebrafish, especially for the auditory system. The corresponding techniques are simple and relatively efficient. Some motor actions from zebrafish are inherent and robust, with no interference from studying and conditioning. Furthermore, these actions have definite characteristics and thus are easy to distinguish [
].
Rheotaxis
The lateral line system along the torso portion of zebrafish comprises nine neuromasts on each side of the body [
]. The major function of these neuromasts is to detect water movement; this function is essential for the ability of fish to escape from predators, detect preys, and socially interact. These abilities are closely related to and can be experimentally evaluated by one of the stereotypical swimming behavior of the fish, namely, rheotaxis. Rheotactic behavior is described as how the zebrafish invariably orients itself toward the water source and swims parallel along a medium-high strength flow to reduce resistance.
Hair cells, aggregated in the center of each lateral line neuromasts, are the essential neurosensory components that mediate the rheotactic behavior. These cells are structurally similar and isogeneous to the mammalian hair cells within the inner ear. The superficial location of lateral line and stereotypical arrangement of neuromasts facilitate pharmacological interventions, as well as morphological detection and comparison. Thus, the function of zebrafish lateral line, or more specifically the rheotaxis behavior, offers an excellent experimental model to study inner ear dysfunction, protection, and repair mechanisms. The severity of rheotactic impairment is closely correlated to the dosage of ototoxic and otoprotective drugs in the aquatic environment; this behavioral response correlates to anatomy-dose response for the same drugs [
]. Therefore, rheotaxis is authentic to assess the function of hair cells.
Stable water flow with appropriate strength is needed to properly measure rheotaxis. Fishes swim freely in low-strength flow and are rushed away in high-strength flow. A stable laminar flow helps accurately quantify the strength of flow. However in early time, non-laminar flow was used in experimental settings. For instance, McNeil
et al. [
] generated a unidirectional vortex flow within a 50 ml beaker using a magnetic stirring apparatus. Four zebrafish larvae were transferred into the beaker and were allowed 2 min to acclimatize to the environment and flow before recording. In the small beaker, the strength of the curve flow varied in different depths and distances to the vortex center. To straighten the curved flow, Olszewski
et al. [
] built a suction chamber for the rheotactic measurement (Fig. 2A). The suction chamber was a cylindrical tank that measures at 8.3 cm in diameter and 7.7 cm in height. The center of the bottom was connected to a suction pump and a fluid container through a tubule. Thus, an inward-radial flow was generated, and the strength of the flow was set to avoid inhaling the zebrafish into the tubule. The suction chamber was submerged in a larger container to supplement water and produce constant suction. This apparatus presented noticeable improvement compared with the McNeil model but still had several limitations. First, the flow was not exactly horizontal and sufficiently stable. Second, the strength of the flow varied at different positions. As a result, in the assay conducted by Olszewski [
], instead of remaining in one position, the zebrafish typically closed up to the suction center in relatively slow velocity and then burst into swimming to avoid the suction center. In the assay, fish were placed in the water only 1.5 cm away from the suction center without allowance of acclimation, which increased inconsistency in the measurement. However, if the fish are placed in a farther position, then faster total flow and larger water containers are required to maintain the same flow strength that the fish experienced. To solve this particular issue, the apparatus was further improved by Olive
et al. [
] who used a pump to recycle water, offering stable flow for a longer time. A horizontal board 4 mm below the water surface was used to replace the suction chamber and a horizontal suction pipe was placed behind the board to produce horizontal flow. However, the flow was turbulent; the small depth of water above the board aggravated the turbulent and was unfavorable for fish swimming. Eventually, a laminar flow that is a circulatory flow system was created by Suli
et al. (Fig. 2B) [
]. A range of 20–30 zebrafish larvae were placed in a long pipe with metallic window screens on both front and rear sides of the fish. The window screens were treated as collimators and produced laminar flow, which was confirmed by the flow coloration. The water temperature was controlled at 26–28 °C, and the fish were allowed 3 min for acclimation. This apparatus can test only one strain of zebrafish or one experimental condition at a time.
Niihori
et al. [
] built a multi-lane apparatus (Fig. 2C) with each lane similar to that of Suli
et al.; the lanes provided laminar flow and controlled stream velocity at 0.15 cm/s. The flow was created via the gravity of water in a head tank. A pump and a pipe between the head tank, as well as a lower collection tank maintained a constant water level in the head tank to feed the multi-lane system. In this method, the flow strength was more stable than that directly created by a peristaltic pump. However, individual channels were 10.2 cm in height, inconveniently resulting in a potential overlap of fish images. In addition, the whole platform was large, whereas the six channels were still insufficient, considering the numerous genetic mutants and drugs to be screened. Further improvement with maximized throughput is required.
So far, Niihori’s 6-channel platform is a relatively advanced mid-throughput apparatus for rheotaxis measurement; thus, we will provide extra discussion for the said equipment. To perform an experimental assay, a group of 5-dpf zebrafish with a pre-treatment condition, such as a particular dose of an ototoxic chemical, was transferred into each lane with wire mesh at both ends for containment of the fish in the central area. This containment allowed capturing of images by a high-resolution infrared-sensitive video camera mounted below and pointing towards the multi-lane container. The experimental system was temperature-regulated and housed in a grounded metal cabinet to restrict ambient light-based visual input and shielding electromagnetic input (if any) that might influence the zebrafish swimming pattern. Orientation of fish body is a comprehensive process contributed by multiple/sensory organs, in which vision is an important one. For this reason, the vision factor must be eliminated during the rheotaxis test. Uniform visual background leads to a similar result with infrared illumination; however, the latter method was applied in most rheotaxis tests.
To generate behavioral rheotactic data, static images showing fish orientation were first acquired every 5 s after the fish were placed in steady laminar flow. For each individual fish in each static image, an axial alignment was determined by the angle between the body axis and flow direction. Experiments revealed a differentiating criterion of 30 degrees, leading to steady statistical outcome in varied testing conditions [
]. Therefore, an axial alignment within 30° and facing towards the direction of the water current was defined as the positive rheotaxis. On the contrary, an axial alignment greater than 30°, or no swimming movement, was defined as non-rheotaxis (Fig. 3). For each experimental condition, a distribution of axial alignments was obtained across individual fish at multiple time points, resulting in a rheotaxis score. A poor rheotaxis score is>30°, suggesting a reduced ability to perform rheotaxis because of the toxic damage of the lateral line system. In addition, the rheotaxis score was positively correlated to the dosage of tested ototoxin and to the severity of the lateral line damage, which was also confirmed by the anatomical examination on neuromasts.
Rheotaxis is only assessed using still images, although full video footages are obtained. Full video stream contains additional information about the swimming behavior beyond the rheotaxis. Analysis of the full-length videos, preferably in real-time, will provide more maneuvers to investigate the neurological processing because the zebrafish responds to steady or dynamic water current.
Lateral line hair cells are highly specialized to detect water current. The outcome of rheotaxis test does not directly determine the function of hair cells to the acoustic signals, such as sound intensity or sound frequency. However, compounds drugs easily access hair cells, which is one prominent advantage of using zebrafish in pharmaceutical studies. Additionally, zebrafish lack blood-labyrinth barrier (BLB) that exists in the mammalian inner ear. The BLB is functionally similar to the blood-brain barrier that produces cochlear endolymph; however, BLB selectively regulates the permeability to each circulating chemicals to mammalian hair cells that are apically immersed in the endolymph. Thus, we need to consider that oto-protective compounds successfully passing through the rheotaxis screen or by other behavior methods cannot easily cross the BLB and gain access to hair cells to exert oto-protection. Local drug delivery methods, such as intratympanic injection, may resolve the issue. On the contrary, we can take advantage of the lack of BLB in zebrafish and use comparative methods between species to separate BLB-specific pathology in mammals. Nevertheless, the rheotaxis test should be selected as an early-stage or assistant experimental measure.
Startle response
The startle response has some definite and stable traits as described by Kimmel
et al. in 1974 [
]. The startle response comprises two stages: the fish body bends into a characteristic “C” shape away from the intense stimuli within 10 ms, followed by a small reversed curve and fast swimming. This action is intense, rapid, and differs from other kinds of behavior, such as “O”-shaped, or “J”-shaped bending. The O-bend occurs within 100–500 ms after the removal of light as indicated by a slow 180° turn of the body [
], whereas the J-bend is a small and slow turn that begins with a slight tail flip to orient the fish in-line with its food source [
]. The startle response is triggered by acoustic stimuli from 5 dpf and throughout adulthood [
]. As the zebrafish age, they retain their response to the same intensity threshold and frequency range [
]. Given the response stability and age independency, the results from the startle responses provide a reliable assessment on the hearing change and pertinent intervening effect.
During the past 20 years, experimental technologies for stimulating and identifying the startle response were developed with improved experimental precision.
(1) Stimulating manners: The simplest trigger for the startle response is a single tap on the edge of the petri dish that holds the fish [
]; the tapping strength was subjectively applied by the researcher. Later, a push solenoid was used to quantify the stimulating strength [
]. Another study used a speaker placed in the air without touching the fish tank to generate pure-tone stimuli [
,
]. Considering that the sound wave was reflected and the intensity was greatly reduced at the air-water interface, a revised approach positioned the speaker in the water [
], but the system was more vulnerable because of submerging. Another refinement involved a tight connection of a mini-shaker to the bottom of a platform beneath the fish container [
], enabling efficient transmission of sound to the fish.
In these stimulating manners, the startle response is induced by sound rather than by water waves. Although tactile stimuli evoke startle response, the lateral line only detects flows and low frequency waves, whereas the inner ear in zebrafish detects vibrations higher than 100 Hz [
]. From the above studies, trigger taps with controlled strengths produced sound and avoided larger water fluctuations. Speakers/shakers provided stimuli of more than 100 Hz, which are beyond the sensing range of the lateral line. Moreover, the startle response for loud tones was not exhibited by the zebrafish with defective swim bladders or Weberian ossicles in the acoustic conductive pathway [
].
(2) Identifying manners: The traits of the startle response are clear but observations are not easily acquired. The difficulty is that the action is too fast to be seen by the naked eye; the C-bend is completed within 10 ms. Initiation of the startle response was evaluated by the sudden change of fish position in earlier studies [
]; thus, distinguishing individual fish is tricky when more than two fish were inside the container. Later, a high-speed camera was used to record movements [
], making it possible to record the fast C-bend movement. Fish positions were subtracted frame-by-frame and the flash movements were manually confirmed. Automated software was developed to track the position of fish head and detect a C-bend by the rapid change of head orientation [
]. Alternatively, open source software allowed the simple confirmation of the startle response by automatically calculating the lengths of the fish; the startle response was indicated by>50% shortening [
]. However, this software was designed for universal motion analysis and was not specific for zebrafish [
]. The software requires a user to manually select four identifiers on each identifiable fish. Therefore, development of truly automated approaches that are efficient and effective in fish identification is in great demand.
(3) Experimental precision: Improvements in experimental precision are gradual and steady. In the beginning, only qualitative outcome was achieved with tapping stimulus and bare-eye observation. With the introduction of the speakers and video cameras, quantitative measurements were conducted to determine the response threshold in terms of sound intensity and frequency range. An integrated instrument was installed for zebrafish startle testing [
], in which several zebrafish were placed in a multi-well plate and synchronously received single-tone stimuli of various frequencies and pressures (Fig. 4). The stimulus was introduced through a board mounted on a mini-shaker. A digital video camera above the wells captured the fish actions. An accelerometer was used to monitor the platform movement during the stimulus presentation. With this instrument, experiments revealed that the “startle response” thresholds were ~60 dB higher than the hearing threshold in adult fish [
].
To fill the gap and use the startle responses to estimate the true hearing threshold, Bhandiwad
et al. adopted the pre-pulse inhibition as the stimulus to improve experimental precision [
]. Pre-pulse inhibition (PPI) is a physiological phenomenon where the startle response decreases when a weaker stimulus appears 30–500 ms before the target stimulus. The decrement of the startle response is directly proportional to the intensity of the pre-pulse, which does not independently initiate the startle behavior [
]. By measuring the decreased startle response, the weak sound (pre-pulse) is detected and the hearing threshold is determined. PPI is inherited, shows minimal adaptation, and is extensively used in auditory studies of humans and other animals [
–
]. A similar PPI threshold to ABR was found in guinea pigs [
], and a more consistent threshold was detected in Mongolian gerbils [
,
]. For the zebrafish, PPI produces a more sensitive result than the standard acoustic startle response and a value similar to those obtained from AEP [
].
Although quantitative measurement of the startle responses was developed with improving sensitivity, most recent studies [
,
,
] still use primitive methods for data collection, such as tapping for stimulation and observation with bare eyes. A reason for retaining these crude approaches is because the construction of the apparatus within a biological laboratory is difficult due to the combination of electronic, mechanical, and image processing technologies. This finding highlights the demand of establishing detailed standards to assemble the apparatus for startle response measurement.
VOR
VOR is a vestibule-executed neural reflex where the eyes shift in the reverse direction of the moving head to stabilize the pictures on retinas. This reflex is used as a conventional test to monitor the vestibular function. VOR indirectly evaluates the functional status of hearing hair cells based on the specific structures in zebrafish. Unlike mammals, zebrafish does not equip the cochlea. The auditory sensor (saccule) and the vestibular sensor (utricle) precisely possess the same type of sensory hair cells. Therefore, VOR tests can also be used in auditory studies [
].
Several technologies are available for measuring VOR. Visual inspection is subjective and is unable to produce quantitative results. Considering that eye movements are driven by muscles, electrodes are placed around the eye to record the musculo-electrical signals. This electro-oculography is inexpensive but susceptible to muscular tension [
]. The sclera magnetic coil is the most reliable method in VOR measurement. A coil contacting the eye in a magnetic field produces electric currents during its shifts. The coil can record three-dimensional (3D) eye movements with high angular and temporal resolution [
]. Drawbacks include the disturbance to the eye shift, complicated operation, and higher cost. Following the developments in camera and image processing, video-oculography (VOG) directly captures and analyzes videos of eye motions, and is gradually replacing the magnetic coil. VOG is non-invasive, well-tolerated, and has increasing accuracy. To measure the eye movement in larval zebrafish, VOG is practically the only choice due to the tiny size of the fish head and eyes.
Several groups developed experimental setups to measure larval fish VOR with different methods in fixing and stimulating the fish larvae. Moorman
et al. placed the zebrafish larva in a glass capillary tube to limit its body motion. The tube was tilted to trigger the gravity-induced VOR, which was recorded by an ophthalmic microscope and a digital video camera [
]. Easter and Nicola placed the fish on a horizontal platform and manually rotated the platform to evoke the VOR. The eye movement was recorded by a micro-amplifier video camera [
]. Recently, Beck
et al. built a system with motorized platform and computerized eye movement analysis system (Fig. 5A) [
]. In this instrument, the fish was placed on a servo-drive platform where speed and position can be obtained in real-time. Temperature was strictly controlled to ensure the physiological activity of the fish. A CCD camera above the platform was aimed at the fish head and rotated along with the platform. Image processing software extracted the video information about the angle and velocity of the swing eyes. Furthermore, an additional module to measure the optokinetic reflex (OKR) was integrated into this instrument. OKR is another type of eye movement performed by the visual and oculomotor abilities and can be used to verify the motor function of the fish larva in the VOR study. OKR was performed prior to the VOR measurement. A drum with white and black stripes encircling the zebrafish was rotated to evoke eye motions following the stripes. If OKR was normal and VOR was missing, then, the defects were positioned at the vestibular sensor rather than at the later motor part of the VOR pathway.
Mo
et al. implemented a compact system with the fish sample and the digital video system both in a motorized rotary platform, as depicted in Fig. 5B [
]. The VOR module was retained but rearranged in a U-shape to reduce the device size. These two VOR instruments (Fig. 5) were imitated and have increased use in zebrafish vestibular research [
–
].
Results from VOR measurement in zebrafish were inconsistent because of the rapid advance in instrumentation. For instance, VOR was reported at 3 dpf by Easter and Nicola, whereas a later study by Beck
et al. reported that no VOR was evident until 35 dpf [
]. This difference is a result of the illumination method. The former study used natural light that induced visual stimulation while the fish rotated; in contrast, the later study adopted invisible infrared. The conclusion of Beck
et al. was also confirmed by a later study [
]. Considering that the device of Mo
et al. [
] lies in different directions, they observed that VOR before 35 dpf was overlooked when the zebrafish rotated in a horizontal plane but was detected in a vertical rotary plane [
]. The semicircular canals that sense angular acceleration of the rotation matured after 35 dpf [
], whereas the otolith organs that sense linear acceleration and gravity matured after 5 dpf. When the zebrafish rotated in the vertical plane, the intersection angle between gravity and otolith organ changed and elicited VOR. Therefore, the instrumentation and technical development improved our understanding on vestibular sensation in zebrafish, whereas the newly acquired knowledge includes additional requirement on future VOR devices. This finding reveals that the healthy relationship between scientific understanding and instrumentation is a major propelling force in our scientific endeavor.
The current VOR equipment still requires several improvements. The first improvement is to measure the 3D eye movement. The previously discussed two instruments are incapable of measuring 3D eye movements. Eye movements in three directions (horizontal, vertical, and rotary) are usually mixed in response to the activities of the body. The principle on the combination of these movements was determined in primates and rodents [
,
]. However, no equivalent research was conducted on fish. Hence, new technology for 3D VOR measurement is imperative. The second improvement concerns the specimen holder. The two aforementioned designs use agarose or methylcellulose to fix the fish body onto the glass slide, with the eyes rolling freely in a drop of water. The droplet distorts the image and thus affects the calculation of the rolling angles of the eye. In addition, the camera is unable to laterally capture the picture because of the miniature size of the fish eyes and the shielding of the glass slide. In one refinement, larvae were fixed in a glass capillary to allow simultaneous video recording from above and lateral positions [
], which, however, adversely introduced a severe anamorphous effect. Therefore, a special specimen holder is required to allow visualization from multiple orientations. Finally, devices used for the startle response and VOR only measure one or several fish at most at a time. The capacity of parallel processing should be improved as that in the rheotaxis measures.
Outlook for the behavioral methods of zebrafish
High-throughput screening
The key factors for high-throughput screening were already fulfilled by the zebrafish model. The zebrafish is an ideal model for large-scale tests. In addition, image processing provides technical advantages in objective tracking and analysis of behavior [
,
]. Developments in robotics accelerate the progress of high-rate screening, in which every step can be automatically performed, from loading, aligning, imaging, to calculating [
,
]. All these phenomena result in rapid technique maturation of high-throughput screening in zebrafish.
Gene screening
Approximately 50% of cases about declined hearing are rooted in mutated genes [
]. About 100 deafness-related genes were identified and estimated to comprise only about one-third of all deafness-related genes [
]. The other two-thirds of these hearing relevant genes remain unidentified. Current approaches for deafness gene discovery are expensive and time-consuming. The screenings using low-cost and highly reproductive zebrafish confirmed numerous auditory-related genes; some of these screenings were adopted from behavioral methods [
,
,
]. Moreover, considering the fast regenerative capability of zebrafish hair cells, initially manipulating the genes implicated in regeneration is possible and preferable in zebrafish rather than in rodents and primates. An example is the use of the startle response to test the auditory ability of zebrafish and to explore the mechanism of the gene
Atp2b1a implicated in hair cell regeneration [
].
Pharmacological screening
Ototoxicity is a common cause of acquired hearing loss, whereas aminoglycoside antibiotics, macrolide antibiotics, and anti-carcinogens are major contributors. Zebrafish display high susceptibility to a variety of ototoxic chemicals, which is similar to mammals [
]. Zebrafish also have multiple convenient routes to access these compounds, including orally, by injection, or by immersion in compound-laden water. The relatively low mass of zebrafish reduces the amount of test compound compared with other vertebrates. This factor is particularly useful for expensive drugs or rare drugs. The application of microfluidic systems on zebrafish further reduces drug usage, equipment volume, and overall expenditure [
,
].
Novel otoprotective drugs are expected to be discovered. To date, no medicine was approved by the Food and Drug Administration of the United States to treat auditory impairment [
]. Accordingly, building biological models and relevant technologies is crucial to develop therapeutic compounds. A fundamental and prior screening scheme should initially focused on hair cell defects, because this phenomenon causes the highest proportion of auditory dysfunction. Zebrafish hair cells were used in otoprotective screening; several antioxidants, including glutathione, limit or prevent ototoxic damage [
,
]. In addition, tissue transplantation into the zebrafish and genetic engineering provide assistance for the screening of individualized therapeutic schedules [
]. Hence, use of zebrafish is a simple and effective strategy to investigate the potential ototoxicity and to search for novel otoprotective agents.
To date, only rheotaxis test is being used to screen auditory-related compounds in zebrafish. Unlike rheotaxis, startle response and VOR have not yet been selected for drug dose-related research or in large-scale applications. Studies using new methodologies are needed for zebrafish research.
Neural network
The combination of different types of zebrafish behavior provides useful patterns to study its neural network. First, the vast amount of data from numerous individual neurons makes the research difficult and the results ambiguous. However, the tiny zebrafish maintains the basic architecture and functional principle of the neural network, and thus simplifies data acquisition and process. Second, behavioral assays provide comprehensive insights for the pathway from external stimuli to biological reactions. Varying types of characteristic behavior indicate dysfunctions in corresponding sections along the pathway. The formation and mechanism of the neural network can be inferred through a series of disassembled behavioral modules. For instance, varied abnormal manifestations of VOR helped in constructing the model of vestibular neural network [
] and in identifying vestibular damages. Third, the progress in optogenetics offers tools to visualize the target cells, as well as allow the researchers to activate and deactivate individual neurons with particular time/location sequence, which results in observational behavioral end points. Therefore, a combination of behavioral and other powerful approaches in zebrafish will promote our understanding on neural circuitry, decision making, and behavioral execution.
Conclusions
Zebrafish have increasing application in auditory studies. Yet, new technologies for appropriate and feasible hearing assessment are required. Hearing function in zebrafish can be readily and accurately assessed via behavioral tests, such as the startle response, VOR, and rheotaxis. Further development will improve the efficiency of these tests. In combination of high-throughput equipment, behavioral methods in zebrafish will be greatly appreciated by the auditory research society and will be used in gene identification, drug discovery, and neural network analyses.
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(338KB)
