Small Molecule Screening and Toxicity Testing in Early-stage Zebrafish Larvae

Summary

This protocol describes a drug screening procedure in zebrafish larvae 3-7 days post fertilization followed by morphological assessment.

Abstract

Zebrafish have become a prominent model organism for human translational research, toxicity testing of small molecules and environmental pollutants, therapeutic drug discovery, and other biomedical research applications. The presented protocol aims to provide a clear guide for drug screening and toxicity testing in early-stage zebrafish larvae (aged 3-7 days post fertilization, dpf). Several methodologies describing toxicity testing in zebrafish larvae have been published, albeit lacking a standardized protocol for exposure durations, rate of media changes, and morphology scoring, amongst others. We experimentally determined the ideal larvae ages for conducting the experiments, the plate sizes that work best for those ages, the rate of media changes that provide reliable results while limiting wastage of test compound, and a morphology scoring system that is easily quantifiable in a statistical manner. The manuscript describes in detail a protocol for toxicity testing in zebrafish larvae aged between 3 to 7 dpf, aiming to target the gap in drug toxicity testing and taking the first step towards a standardized method similar to the Fish Embryo Acute Toxicity test.

Introduction

Zebrafish (Danio rerio) have recently emerged as a powerful vertebrate model organism, becoming increasingly popular for studies to advance biomedical research¹. Around 70% of human protein-coding genes have at least one zebrafish ortholog, with ~82% of human disease-causing genes having at least one orthologous gene in zebrafish². The genetic conservation, coupled with similar basic organ patterning and morphology, make zebrafish an interesting model for human disease research,allowing for cross-species comparisons at the molecular level^1,2,3, which is crucial for drug development.

Zebrafish develop rapidly with the tail bud forming at 10 h post fertilization (hpf) and the first somites appearing at 16 hpf. The embryos emerge from the chorion at ~2-3 days post fertilization (dpf). By 4 dpf, their digestive tract is fully developed and by 5 dpf, their swim bladder has inflated, allowing them to swim around freely and hunt for food. Despite the ability to feed by 5 dpf, larvae that are unfed can survive solely on yolk-derived nutrients until at least 7 dpf, circumventing the need for external feeding that may cause contamination and interfere with drug screening, toxicology, and behavioral testing⁴. By 7 dpf, larvae also develop a complete body plan with several rudimentary organ systems, including the heart and vasculature, muscle, and bone^5,6. Zebrafish embryos are also fertilized ex vivo, developing externally, and are optically transparent in the early life stages, making them highly accessible for gene editing and screening of small molecules³. Additionally, their small size and high fecundity rate allow for the inclusion of more replicates, which increases statistical power, as well as for the use of multi-well plates, which saves on drug volumes and increases throughput.

Screening of novel small molecules in an animal model is important to determine the drugs' ability to produce both beneficial or adverse effects, thus identifying which compounds are viable for further testing and others that might require modification to alter their activity. As part of the drug screening process, the lethal concentration 50 (LC₅₀) of the target compound/s needs to be identified. This provides a working concentration that causes death in 50% of test animals as part of the preliminary toxicity testing.

A standardized protocol exists for toxicity testing in zebrafish embryos up to 96 h post fertilization (hpf), namely the Fish Embryo Acute Toxicity (FET) test (TG236)⁷. In the FET test, newly fertilized eggs are exposed to five concentrations of a chemical for up to 96 h and observed for signs of lethality, namely (i) egg coagulation, (ii) lack of somite formation, (iii) lack of tail bud detachment from the yolk sac, and (iv) lack of a heartbeat. The test allows for the determination of acute toxicity and the LC₅₀. Furthermore, it also advises on test chamber sizes to use for screening (24-well plates), egg distribution in the test chamber, and housing conditions, amongst others, ensuring valid and reproducible results⁷.

While the FET test is suitable for toxicity testing at the embryonic stage, it does not go beyond 4 dpf. For phenotypes that start to develop at around this time point (e.g., mineralization of the notochord⁶), drug screening must proceed further than 4 dpf to enable the morphological assessment of such tissue. Moreover, we recommend that drug exposure commences at least a day before the development of the phenotype of interest. To our knowledge, no standardized protocol similar to the FET exists for drug screening in early-stage zebrafish larvae, specifically in zebrafish aged from 3 to 7 dpf. While several studies have been published describing drug testing in zebrafish larvae beyond 4 dpf, there is a lack of consistency in the methodology utilized, including the size of the test chambers, number of biological and technical replicates, rate of media changes, exposure durations, and morphological defects assessed^{8,9,10,11,12,13}. For this reason, we developed a validated protocol for standardized toxicity testing followed by visual observation of morphological changes in zebrafish larvae at 3-7 dpf. The proposed protocol can subsequently be utilized for other downstream phenotyping assays, such as behavioral analysis, gene expression profiling, and histomorphometry.

Protocol

Zebrafish were handled according to EU Directive 2010/63/EU. All experiments were approved by the Faculty Research Ethics Committee (FREC) and the Joint FREC Animal Research Sectoral Subcommittee (JFARSS) of the University of Malta. A flowchart of the proposed protocol can be found in Figure 1.

1. Egg production, collection, and maintenance

1. Set up the fish in breeding tanks the day before, at least an hour after the last feed, with dividers in place separating the males and females. For better egg production results, use a higher ratio of females to males. Set the fish up in a room equipped with a 14 h:10 h light/dark cycle and at 27-28 °C ambient temperature.
   NOTE: A 1 L breeding tank can hold three adult fish, whereas a 1.7 L breeding tank can hold five adult fish.
2. The following morning, following the onset of light, replace the water in the breeding tanks and remove the dividers. In case the breeding tank is not readily sloped, place it in a sloping position to ensure that shallow water is available to promote mating. In case of slow egg production and mating activity, replace the water in the breeding tanks every 45 min. Leave the fish to mate for up to 3 h or as required.
3. Once mating is complete, gently squeeze any females that did not produce eggs to prevent long-term egg binding. Place the fish back into their respective tanks.
4. Collect any eggs in the breeding tanks using a mesh strainer, rinse with embryo water (E3), and place in a Petri dish with E3. To prepare E3, mix 16.67 mL of 60x E3 (300 mM NaCl; 10 mM KCl; 20 mM CaCl₂.2H₂O; 20 mM MgSO₄.7H₂O) with 1 mL of 0.01% methylene blue and 982 mL of deionized water. Remove any unfertilized eggs or debris. Make sure the Petri dishes are not crowded, with 50 eggs per 100 mm x 15 mm Petri dish and 120 eggs per 150 mm x 15 mm Petri dish.
5. Keep the eggs in an incubator set to 28.5 °C and equipped with a 14 h:10 h day/night cycle. At 1 dpf, refresh the E3 and remove non-viable eggs to limit fungal growth.

2. Test plate preparation

1. At 3 dpf, transfer one healthy-looking, hatched embryo to each well in a 96-well plate using a 1,000 µL micropipette set to a 100 µL volume.
   NOTE: Non-hatched embryos should not be used, as this may result in variable compound uptake. Using embryos that appear healthy also limits bias during testing. The size of the pipette tip allows for easy collection of the zebrafish embryo without causing any physical damage. In addition, a micropipette allows for better control of the volume being dispensed into the well as opposed to a transfer pipette.
2. In a 12-well plate, prepare the desired concentration(s) for each test chemical in separate wells (here, Bosutinib at a concentration range from 100 µM to 6.25 µM). In each well, add the necessary volume of test chemical using micropipettes and make up to 2.4 mL using E3.
   NOTE: A literature search should be carried out to determine the adequate concentrations to use for the relevant test chemicals. If no information is available, several concentrations can be trialed until a suitable concentration range is determined. A minimum of five concentrations should be used, with the highest concentration causing 100% mortality and the lowest concentration causing no observable effect⁷.
3. Prepare a vehicle control depending on the solvent used to dissolve the stock test chemical.
   NOTE: Many test compounds are not water soluble and for this reason, they need to be dissolved in specific solvents ("vehicles") such as ethanol or dimethyl sulfoxide (DMSO) to generate an aqueous solution. DMSO should preferably not exceed a concentration of 1% when used as a vehicle control, which is the concentration found to be safe in zebrafish¹⁴.
4. Remove any E3 solution present in the 96-well plate using a 200 µL micropipette and replace it with 100 µL of the corresponding medium being tested. Be sure to test a total of 24 larvae per concentration. Use a single or 8-channel 200 µL micropipette to transfer the test chemicals to the corresponding test plate wells. Include controls (e.g., E3, chemical vehicle, positive controls, etc).
   NOTE: The plate setup accommodates up to four different test chemicals or four different concentrations of the same chemical. An example of the test plates set up for the investigation of five concentrations of the same chemical (e.g., Bosutinib) accompanied by controls is shown in Figure 1.
5. At 24 h post exposure, perform morphological scoring (as described in section 3) and a halfmedium change, removing 50 µL of medium and replacing it with freshly prepared medium. Remove and record any dead larvae in the process. Keep the larvae in an incubator set to 28.5 °C and equipped with a 14 h:10 h day/night cycle. Repeat the process until 7 dpf.

3. Morphological assessment

1. Using a stereo light microscope, observe the individual test larvae at an appropriate magnification.
   NOTE: Examples of morphological defects that can be assessed are provided in Table 1.
2. Have the morphology assessed by two independent scorers, with a third one being consulted in case of disagreements between the two scorers.
3. Score the morphology using a binary method (i.e., present/absent), where 'absent' indicates normal morphology (scored as '0') and 'present' indicates the presence of morphological abnormality (scored as '1'). Figure 2 shows examples of zebrafish at 5 dpf after 2 days of drug exposure, including an unaffected zebrafish (Figure 2A), a zebrafish with mild pericardial edema (Figure 2B), severe pericardial edema with yolk hemorrhage and yolk extension (Figure 2C), delayed development with reduced pigmentation, swim bladder and total length (Figure 2D), swollen yolk and a kinked caudal fin (Figure 2E), absent caudal fin (Figure 2F), curved notochord (Figure 2G), truncation (Figure 2H) and death with necrosis (Figure 2I). Score all abnormal phenotypes as '1' and the unaffected zebrafish as '0'.

4. LC₅₀ determination

1. Following morphology assessment, plot the cumulative percentage mortality at 7 dpf against concentration. An example is shown in Figure 3. Use the concentration curve to estimate the concentration that causes 50% mortality.
2. Once the LC₅₀ concentration is estimated, repeat the experiment described above by exposing 24 replicate larvae in a 96-well plate to this single estimated LC₅₀ concentration.

5. Technical replicates

1. Repeat the entire procedure using the selected five concentrations of the same chemical at least two more times on two separate occasions to ensure that the results are reproducible.

Results

The abnormalities can be looked at individually or cumulatively over the experimental timeline until 7 dpf. The maximum morphology score per abnormality can be used to calculate morphological abnormalities as a percentage of the number of larvae used. For example, if on day 7 the pericardial edema morphology scores are at 19 and 20 for scorers 1 and 2, respectively, out of a possible score of 24, meaning 19 or 20 larvae out of 24 display pericardial edema, the percentage score can be calculated as:

Supplementary Table S1 shows the morphology scoring results (of both scorers) at 4 dpf following toxicity testing for the compound Bosutinib, a tyrosine kinase inhibitor indicated for the treatment of chronic myeloid leukemia¹⁵. The same morphology scoring is repeated for exposed fish aged from 5 dpf to 7 dpf. The compound was tested using a concentration range from 100 µM to 6.25 µM. Three technical replicates were conducted at different time points, and the average cumulative mortality rate at 7 dpf was used to determine LC₅₀, the concentration at which 50% of the larvae survive. The LC₅₀ can be determined at any exposure time point. However, we recommend calculating this at 7 dpf.

The cumulative average mortality at 7 dpf was used to generate an LC₅₀ curve, which can be seen in Figure 3. For Bosutinib, the LC₅₀ concentration was estimated to be 37.95 µM.

NOTE: This estimated LC₅₀ concentration strictly applies to this model system and can then be used for further downstream phenotyping assays.

Figure 1: Schematic of the proposed drug screening in larvae at 3-7 dpf. The 96-well test plates are set up for the investigation of five concentrations ranging from 100 µM to 6.25 µM for the compound Bosutinib. Abbreviation: dpf = days post fertilization. Please click here to view a larger version of this figure.

Figure 2: Representative morphological defects observed in zebrafish larvae at 5 dpf exposed to various chemical concentrations (at 2 days post exposure). (A) Control (unaffected) zebrafish, (B) Zebrafish larva showing mild pericardial edema, (C) Zebrafish with severe pericardial edema and yolk hemorrhage, (D) Zebrafish with developmental delay, reduced pigmentation, small swim bladder, (E) Zebrafish with swollen yolk and a kinked caudal fin, (F) Zebrafish with absence of caudal fin, (G) Zebrafish with curved notochord, (H) Zebrafish with truncated tail, and (I) Dead zebrafish with necrosis. Defects are noted with red arrowheads. Scale bars = 1 mm. Please click here to view a larger version of this figure.

Figure 3: LC₅₀ curve showing concentration against % mortality. The five concentrations tested were 100 µM, 50 µM, 25 µM, 12.5 µM, and 6.25 µM, with % mortality being 100%, 100%, 29.2%, 12.5%, and 12.5%, respectively. Please click here to view a larger version of this figure.

Morphological Defects	Description
Pericardial edema	Fluid buildup around the heart with a stretched heart and malfunctioning of the blood circulation
Heart beats (per minute)	Reduced number of beats per minute
Yolk	Presence or absence of yolk edema or hemorrhage that can occur simultaneously with cardiac defects
Yolk extension	Yolk which is swollen, thin or absent
Curved notochord	Presence of a kinked or undulating notochord, and/or curved tail
Swim bladder	Presence, absence or partial development of the swim bladder
Pigmentation	Less or no melanocytes compared to the untreated fish of the same age
Necrosis	Presence (mild or severe) or absence of necrosis in any part of the body
Delayed development	Delayed development or stalled compared to wild-type controls of the same age.
Response to touch	Delayed or no response following a startle (e.g., using a pipette tip)
Truncation	Absence of posterior zebrafish body (i.e. formation of the tail)
Fins	Presence or absence of the caudal fin that may be accompanied by thin yolk extension
Reduced body axis	Fish that appear normal but when measured are shorter in length

Table 1: A list of visual morphological defects that can be observed following drug exposure studies.

Supplementary Table S1: Representative morphological results for Bosutinib at 4 dpf (1 day post exposure). Please click here to download this table.

Discussion

Zebrafish have become an important model for drug screening. When trying to determine a methodology to carry out toxicity testing, we found several contrasting methods from different laboratories^{8,9,10,11,12,13}. The methods described use larvae at different ages, variable experimental timelines, different rates of medium changes, and different morphology scoring systems. Through trial and error, we established the described methodology, which has provided reliable and reproducible results.

Choosing a plate size
Initially, experiments were performed in a 24-well plate to ensure that the larvae had enough space to move and grow. The plates were arranged with 10 larvae per well, two wells per concentration, and 1 mL of drug solution per well. This arrangement did not provide reproducible results, leading to the inference that the larvae were not absorbing equal amounts of drug from the well. Trials were then carried out using a 96-well and 48-well plate with one larva per well, which showed that larvae in 96-well plates displayed no issues with growth. Hence, experiments were continued in a 96-well plate utilizing less space and drug volumes.

Daily medium change
Since zebrafish are absorbing the drug from the medium, we ascertained that a degree of medium change is needed to keep a consistent concentration of the drug in the medium. Moreover, a medium change removes potential waste material accumulated in the well, which can provide a toxic environment and lead to biased results. A trial was carried out with Bosutinib, where three plates were set up for the same drug, one without a daily medium change, one with a daily half medium change, and one with a full daily medium change. While the plate with no medium change exhibited increased mortality rates, the plates with half and full medium changes resulted in the same outcome by 7 dpf. Thus, a daily half medium change was chosen. This choice ensures less consumption of the expensive and limited test material, which could be required for further downstream experiments.

Exposure duration
The age of test larvae used and drug exposure duration selected in this protocol was driven by our phenotype of interest being bone mineralization. Since bone formation commences at ~4 dpf, we chose to start with zebrafish at 3 dpf to determine whether test chemicals affected the onset of bone development. Subsequently, we ended the drug exposure studies at 7 dpf since a high degree of mortality was observed from 8 dpf onwards in view of the lack of feeding that might bias the results. Nevertheless, timeframes may be adjusted depending on the nature of the study and phenotype of interest.

Limitations
This methodology has some limitations. Transferring 3 dpf zebrafish to multi-well test plates is time-consuming. Removing the medium from the wells without picking up the larvae is equally laborious. The medium needs to be removed from each well slowly using a single-channel pipette, ensuring that the larvae are not damaged in the process. Moreover, while a daily half medium change replenishes the test compound and ensures the removal of waste material from the well, it does not guarantee that a stable drug concentration can be maintained in the well throughout the 5 days of exposure. While high-performance liquid chromatography analysis could give further insight into the actual amount of drug being absorbed by each larva, the test would no longer qualify as a high-throughput screening method.

In conclusion, this methodology aims to target the gap in toxicity testing using zebrafish beyond 4 dpf, taking steps towards a standardized protocol applicable to wild-type, mutant, and transgenic zebrafish lines, and that is on par with the FET at the embryonic stage. This opens avenues for other downstream phenotyping assays that might not be possible at earlier stages.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
1.5 mL microcentrifuge tube	Starlab	E1415-1510
10 µL micropipette (P10)	Gilson	F144802
10/20 µL pipette tips	Starlab	S1110-3810
1000 µL micropipette (P1000)	Gilson	F123602
1000 µL pipette tips	Starlab	S1111-6001
12-well plate	Starlab	CC7672-7512	Sterile, single-wrapped, non-treated, with lid
20 µL micropipette (P20)	Gilson	F123600
200 µL micropipette (P200)	Gilson	F123601
200 µL multi-channel pipette	Gilson	F81024
200 µL pipette tips	Starlab	S1113-1006
96-well plate	Starlab	CC7672-7596	Sterile, single-wrapped, flat-bottom, non-treated, with lid
Breeding tanks with inside grid sloping bottom, tank dividers and lids	Tecniplast	ZB17BTISLOP, ZB17BTE,  ZB17BTL, ZB17BTD	Sloped inner tank
Mesh strainer	/	/
Nitrile gloves	Mercator	/	/
Petri-dish	Starlab	CC7672-3394/CC7672-3359	100 x 20 mm or 60 x 15 mm
Leica M205 FCA Fluorescence stereo microscope	Leica	10450826
Peltier-cooled incubator including 2 shelves with light module cold white	Memmert	IPP110ecoplus, T8	Equipped with day/night cycle
Dimethyl Sulfoxide	Biochem Chemopharma	504341000	To prepare a 1% solution in E3
Calcium chlorided dihydrate	Biochem Chemopharma	303080500	To prepare E3
Magnesium sulfate heptahydrate	Biochem Chemopharma	313060500	To prepare E3
Potassium Chloride	Biochem Chemopharma	316030500	To prepare E3
SCREEN-WELL Wnt Pathway library (Bosutinib)	Enzo	BML-2838-0500
Sodium Chloride	Fisher Scientific	S/3161/53	To prepare E3