Materials. AB strain zebrafish (supplied by Zebrafish Research Center of Shandong First Medical University), zebrafish culture system (Shanghai Haisheng Biological Experimental Equipment Co., Ltd.), and pentylenetetrazole (PTZ, Sigma-Aldrich) were sourced. Kainic acid (KA, Lot: C12869133, Purity: ≥99%), aminophylline (AMI, Lot: C10308420, purity: ≥99%), sodium valproate (VPA, Lot: C11370830, purity: 98%), phenytoin sodium (PHT, Lot: C12637040, Purity: 99%), topiramate (TPR, Lot: C11930907, Purity: 98%), carbamazepine (CBZ, Lot: C12698337, Purity: 99%), ethosuximide (ESM, batch number: C11882547, purity: 98%), D-sorbitol (SOR, batch number: purity: 98%), rifampicin (RFP, C12829015, 97%), the above reagents purchased from Shanghai McLean Biochemical Technology Ltd. SparkZol Reagent extraction reagent, SPARKscript II RT Plus Kit (With gDNA Eraser) reverse transcription kit, 2×SYBR Green qPCR Mix (With ROX) qPCR kit, MiNi Vortex (DL-SC05) purchased from Beijing Donglin Changsheng Biotechnology Technology Co., Ltd. Sony video camera (DSC-RX100M7, Sony Corporation, Japan), MATLAB software (MathWorks, USA) and MARGO software package (https://github.com/de-Bivort-Lab/margo). The VELOCITY 18R high-speed refrigerated centrifuge (Dynamica), the PowerPac Basic electrophoresis instrument (Bio-Rad), JY300C electrophoresis tank (Beijing Junyi Dongfang Electrophoresis Equipment Co., Ltd.), the Tanon 1600 Gel Imager (Shanghai Tianneng Technology Co., Ltd.), the Nano-300 microspectrophotometer (Hangzhou Aosheng Instrument Company), Roche LightCycler 480 real-time quantitative PCR instrument (Roche Company).
Zebrafish maintenance and egg collection. AB strain adult zebrafish were cultured under a constant temperature culture system at 28 ± 0.5°C in a 14-h/10-h light/dark cycle (light at 08:30 AM and dark at 10:30 PM). Adult zebrafish were fed Artemia twice a day, and male and female zebrafish were placed on either side of the partition of the mating tank in a ratio of 2:2 at 09:00 PM. The next day, pull out the partition, and zebrafish would mate and spawn after being stimulated by light. The eggs were collected within 2 h of spawning and rinsed thrice with dedicated fish water (5 mM NaCl, 0.17 mM KCl, 0.4 mM CaCl2, and 0.16 mM MgSO4), and the clean embryos were transferred into a 90-mm petri dish containing fish water. The unfertilized and stunted eggs were removed, and the remaining eggs were incubated at 28.5°C for subsequent analysis. All animal experiments were reviewed and approved by the Ethics Committee of Shandong First Medical University (Shandong Academy of Medical Sciences). All experiments were in accordance with the regulations of Shandong First Medical University (Shandong Academy of Medical Sciences). And all experiments conformed to the ARRIVE guidelines for involvement of animals (fish).
A new algorithm to identify the swirly swimming behavior of zebrafish. (1) Video capture. A zebrafish behavior acquisition device was designed for this experiment, consisting of a square box closed around (40cm × 40cm ×40-cm dimension), and a bottom transparent baseplate with LED light that a microplate can be put on it. A high-resolution camera (4K) was placed directly above this setup. The 5-day post-fertilization (5dpf) larval zebrafish were placed in the microplate, and the camera focus was adjusted until the camera field of view was moderate and the larvae could be seen clearly. After the set-up was ready, the swimming state of the larval zebrafish was photographed for 30min (Fig. 1A).
(2) Image preprocessing. Image preprocessing refers to the use of digital technology to eliminate useless information in an image and enhance the detectability of the desired image, thereby improving the accuracy and reliability of feature extraction, image segmentation, and target detection. The captured video was imported into MATLAB (MathWorks, USA) and MARGO (https://github.com/de-Bivort-Lab/margo). After the tracking area was set, MARGO would automatically convert the video into grayscale frame-by-frame images. Then an appropriate threshold value was set. If the gray value is greater than or equal to the threshold value, it will be the detection target, otherwise it will be the background image. The processing formula is shown as follows:
\(G\left(x,y\right)=\left\{\begin{array}{c}0,f\left(x,y\right)<A\\ 1,f\left(x,y\right)\ge A\end{array}\right.\) , (1).
In the formula, \(G\left(x,y\right)\)represents the gray value of the binary image, \(f\left(x,y\right)\)represents the gray value of the current frame image pixel, and A represents the set threshold. The accuracy of target detection was improved by adjusting the threshold value to remove the noise in the image.
The target zebrafish and the background were segmented via background subtraction for target recognition. The process of background subtraction was expressed as follows:
\({D}_{k}\left(\chi ,y\right)=\left|{f}_{k}\left(x,y\right)-{b}_{k}\left(x,y\right)\right|\) , (2).
\({N}_{k}\left(x,y\right)=\left\{\begin{array}{c}1{ forground D}_{k}\left(x,y\right)>A\\ 0 background {D}_{k}\left(x,y\right)\le A\end{array}\right.\) , (3).
In the formula, \({D}_{k}\left(\chi ,y\right)\) represents the gray value of the differential image, \({f}_{k}\left(x,y\right)\) represents the gray value of the selected current-frame image, \({b}_{k}\left(x,y\right)\) represents the gray value of the background image, and \({N}_{k}\left(x,y\right)\) represents the gray value of the binarized image. Please see Fig. 1B.
(3) Calculation of the vector angle. MARGO can output 25 frames of zebrafish swimming moments and the coordinates of each object in 1 s. And the swimming frames and 2D coordinates were used to calculate the zebrafish swimming distance and the angle between the vectors within a certain period (Fig. 1C). The calculation formula 14 for the vector angle is depicted below:
\(\left(\theta \right)=\frac{\left(\left({{tan}}^{-1}\left(\frac{\varDelta Y}{\varDelta X}\right)\right)*180\right)}{\pi }\) , (4).
In the formula, \(\theta\) represents the included angle of the vector, plus or minus represents the direction, Δ X, Δ Y is the difference between the horizontal and vertical coordinates of the next frame and the current frame, respectively.
(4) Extraction of the seizure behavior parameters. The swimming process of zebrafish can be described based on the following parameters: the vector angle θ at the moment of swimming, the trajectory length s in a certain period, and the swimming duration t. A normal zebrafish can hardly swim with abrupt turning except under pathological stimulus, while PTZ-treated zebrafish would produce distinct swirly swimming. After frame-by-frame observation, it was found that the average time of a vortex was 0.4s. Therefore, under the present experimental condition, a pathological whirlpool-like swimming behavior of larval zebrafish needs to meet simultaneously the following conditions: (1) the zebrafish completes more than 3/4th of a circle in a limited time, (2) the zebrafish performs 1 cycle during the swirly swimming, the vector angle θ > 30° in 2 consecutive frames, (3) the number of co-directional angles >30° between 25 consecutive frames n > 12, the trajectory length s > 7.5 mm, and the swimming duration was t > 0.3 s (because zebrafish would not generally stop during a swirly swimming process). The total number of the swirly swimming within 30 min was defined as N, and the time when the swirly swimming appeared for the first time was defined as the seizure latency T. By setting the abovementioned parameters, an innovative behavioral analysis system was established (Fig. 1D).
PTZ induces behavioral changes in zebrafish. The larval zebrafish was placed in a microplate and treated with 2.90-mM PTZ for 30 min, with a blank control group set at the same time. Using the camera to capture 30 min swimming state video of each group of zebrafish, and then import them into the analysis system.The size range of the vector angles included θ >30°, >40°, >50°, and the swimming duration t was 0.3 s, 0.4 s, and 0.5 s, forming parameter combinations (30, 0.3), (30, 0.4), (30, 0.5), (40, 0.3), (40, 0.4), (40, 0.5), (50, 0.3), (50, 0.4), and (50, 0.5). The combination of the above-mentioned parameters was written into the code to output the swirly characteristic data of zebrafish (the number of swirly swimming N, the seizure latency T), and the difference between the blank groups of the model groups was recorded.
Comparison of the results from the output of the analysis system and those with naked eyes for counting frame-by-frame. The larval zebrafish were treated with 1.45, 2.90, 4.35, and 5.80 mM PTZ solutions, and cameras were used to capture simultaneously the video of zebrafish in all treatment groups. The captured videos were imported into the analysis system for tracking and outputting results of seizure behavior in larval zebrafish. A 30-min video of the swimming behavior of individual was read with the open-source video playback software. Meanwhile, three people separately read the video frame-by-frame using a double-blinded method. The data obtained by these two methods were subjected to statistical analyses.
Testing for seizure-associated genes. The larval zebrafish were treated with 1.45, 2.90, 4.35, and 5.80 mM PTZ for 30 min, while a blank control group was set up. In the end, 60 zebrafish from each group were collected, washed thrice with double-distilled water, and placed in a 1.5-mL centrifuge tube. Next, the culture water was removed from the tubes and immediately frozen in liquid nitrogen. These operations were repeated thrice, and the frozen zebrafish were disrupted, followed by extraction of the total RNA with the SparkZol Reagan reagent, and reverse transcribed into cDNA by using the SPARKscript Ⅱ RT Plus Kit (With gDNA Eraser). Real-time PCR reactions were performed with the Roche LightCycler 480 real-time quantitative PCR instrument, and the 2×SYBR Green qPCR Mix (With ROX) was used to quantify the gene expression. Relative quantification was performed by using the 2−ΔΔCt method 15 with GAPDH serving as an internal reference 16. Forward and reverse primers were designed as follows, GAPDH-F 5′-TGACCCGTGCTTTCTTGAC-3′, GAPDH-R 5′-TTGCCGCCTTCCCTTAACC-3′, c-fos-F 5′-AGCAGACGAGAGAGAGGAATAC-3′, and c-fos-R 5′- TCGGCGGTTAAGGCTGGTAAACATC-3′.
Correlation of the vortex number output from the ZebVortrack system with the relative expression of c-fos was analyzed using Pearson Correlation Analysis.
Different pro-convulsant agents induce seizures in zebrafish. Different concentrations of PTZ (1.45, 2.90, 4.35, and 5.80 mM), AMI (2.4, 4.8, 9.5, and 14.3 mM), KA (86.5, 130.0, 173.0, and 216.0 µM), and a blank control group were prepared. Each swimming video of 30 min was imported into the ZebVortrack system for zebrafish behavior analyses.
Effects of different antiseizure drugs on the seizure behavior of PTZ-induced larval zebrafish. Antiseizure drugs VPA (360, 480, and 600 µM), ESM (1.77, 3.55 and 7.10 µM), PHT (36.5, 73 and 109 µM), TPR (30, 45 and 60 µM), CBZ (42, 84, and 126 µM), negative-control drugs antibiotic rifampicin (73, 97 and 122 µM) and diuretics SOR (44, 220 and 660 µM) were used to pre-treated 4dpf zebrafish for 24 h, followed by treatment with the 2.90-mM PTZ solution to induce epileptic seizures. Meanwhile, the 2.90-mM PTZ model group and the blank group were set up for comparison. Each swimming video of 30 min was imported into the ZebVortrack system for behavioral analysis of zebrafish.