Robyn Tanguay “vividly recall[s] memories of isolation at Society of Toxicology meetings in the 1990s,” as she stated in an editorial for Toxicological Sciences (1). “Rodent systems predominated, and the appetite for new models such as zebrafish was scant.”
“It was kind of a disruptive technology,” said Tanguay, a molecular toxicologist at Oregon State University who pioneered toxicology studies in zebrafish. “We were a threat to the status quo of in vivo data.”
At the same time, another disruptive technology emerged in the form of nanomedicine, in which nanomaterials on the order of one billionth of a meter in size are used in drug design. With their tiny sizes and tunable surface properties, nanomaterials can easily sneak into cells and across biological membranes, making them especially attractive for drug delivery and diagnostic applications. The number of scientific papers reporting advances in nanomedicine has skyrocketed in the turn of the century, yielding more than 50 FDA-approved nanodrugs since the first in 1995 (2). Recently, mRNA vaccines for COVID-19 relied on encapsulation in lipid nanoparticles to prevent charge repulsion as they cross the cell membrane and to protect the mRNA from nuclease degradation (3).
“Nanomedicine is really making a difference here in the world,” said Stefan Wilhelm, a nanomedicine researcher at the University of Oklahoma. “I'm sure there is a lot that we will hear about nanomedicine and the success stories. But with every new technology, there are always challenges. So, we also need to make sure that as a community, we start to address challenges that have to do with adverse effects, toxicity, safety, and so forth.”
Scientists have come to appreciate that not all nanomaterials are inert bystanders in biological systems. To unravel the relationships between nanomaterial design and toxicity, Tanguay and other researchers have doubled the disruption, evaluating the effects of these materials on zebrafish development, physiology, and behavior. By leveraging a model that bridges the efficiency of cell culture with the complexity of animals, they aim to guide the development of safe nanodrugs.
The ultimate biosensor
When it comes to testing nanomaterials for adverse effects, the advantages of zebrafish are literally quite clear. Before they develop their namesake’s signature stripes, early life stage zebrafish are optically transparent, enabling scientists to easily visualize their internal physiology. They also develop from caviar-scale embryos into larvae approximately four millimeters in length in a matter of days, allowing researchers to rapidly test small quantities of nanomaterials across various life stages.
To expose zebrafish embryos, Tanguay’s team adds nanomaterials to the surrounding water, “which is honestly no different than a cell culture exposure,” she said. “These materials are actually crossing through epithelial membranes, just like they would be going through membranes in most cells.”
If these materials have a surprise interference, for example, with some other process that you can't model in the individual cell, at least you’d see it.
- Robyn Tanguay, Oregon State University
They opt to remove the chorion, the protective shell that envelops the embryo, to eliminate the possibility that a nanomaterial doesn’t show toxicity simply because it cannot pass through. “That’s been a hotly debated topic,” Tanguay said. “We've argued for a long time that we don't know when the chorion might be a barrier. …Since we often do our studies where we're blinded to the identity of the material, we just thought it was more prudent to take the chorion off.”
While nanomaterial exposure studies are conducted similarly in cells and zebrafish, the fish reveal a lot more than just cytotoxicity. “If these materials have a surprise interference, for example, with some other process that you can't model in the individual cell, at least you’d see it,”Tanguay said. For example, the researchers monitor the nanomaterial’s effects on the embryo’s development, looking for mortality or morphological defects in their organs and other bodily structures. Disruption of any of the biological pathways that are carefully choreographed to create a healthy organism will manifest during embryonic development, providing “the ultimate biosensor” for toxicity, Tanguay explained. “The genes that are necessary to build all the components and the body parts of the zebrafish are really similar to the ones that build a human or a mouse or a frog,” she said. “If you're going to use any nonhuman model to predict the toxicity of anything, the best time to do that would be during embryonic development because that's the most complex period of time with the most conserved mechanisms.”
The researchers can also follow the effects of the nanomaterials on the zebrafish’s behavior as the embryos blossom into mobile larvae. They use stimuli like switching between light and darkness or blasting a loud noise to startle the larvae and look for changes in their movement. Impaired motor activity could result from effects on neurological, muscular, sensory, or other systems. “The reason we do behavior [tests] is to add that sensitivity to that biosensor,” Tanguay said. “It really increases the sensitivity to see other evidence that an exposure affected normal processes.”
A nanoparticle that causes blindness in zebrafish doesn’t necessarily translate to an effect on vision in humans, however. “Often, people try to map the effects of these materials across species, and I think that usually doesn't work,” Tanguay said. “What we really want to do is to discover adversity. …If we see an adverse outcome in zebrafish, it makes you want to take a closer look to try to understand what was modulated through those exposures, what targets were hit, what are the functions of that target, and then look for those in other species.”
Tanguay’s team has discovered signs of adversity in certain gold nanoparticles, which have been attached to drugs to enhance delivery, explored as radiotherapy and phototherapy agents for cancer, and harnessed for bioimaging applications (4). The researchers tested a series of gold nanoparticles decorated with different mixtures of molecules on their surfaces. They observed that particles coated with a higher portion of detachable, weakly bound molecules showed more toxic effects on zebrafish embryo development and larval motor activity (5). In a related study in which the team assessed the toxicity of gold nanoparticles of various sizes at a range of concentrations, they observed the strongest correlation between a higher total gold surface area inside a well and greater toxicity toward the exposed zebrafish (6). These results suggest that the nanoparticles’ toxicity stems from having more surface area available to react with proteins, lipids, or other biomolecules.
The researchers have also exposed zebrafish to graphene oxide nanomaterials, which have been investigated as a platform to selectively deliver the heart-damaging chemotherapy drug doxorubicin to cancer cells (7). They observed that an oxidized graphene oxide nanomaterial had lethal effects on zebrafish embryos, but that mortality was significantly attenuated when they added fetal bovine serum (FBS), a growth supplement used in cell culture media (8). FBS proteins can coat the surface of the nanomaterial (forming a protein crown or “corona"), which might mask its toxic effects and produce a false negative result in cell culture studies. “We've found that the zebrafish model is really versatile in the sense that we can go from no proteins and no constituents and then start adding that back to actually control the protein corona,” said Lisa Truong, a toxicologist at Oregon State University who works with Tanguay. Some have argued that this serum shielding effect may more closely represent a nanomaterial’s fate inside a human. “As scientists, our responsibility isn’t to say only the best case scenario. We need to see what happens when it’s the direct contact of that material,” said Truong.
Another potential confounder in cell culture studies is that nanomaterials tend to aggregate in the presence of salt, forming clumps that can settle at the bottom and influence toxic interactions. Tanguay’s team found that early life stage zebrafish tolerate a wide range of ionic strengths, enabling them to manipulate the conditions of the water. “When we took away the salts and kept that material in suspension, then we could actually look at the intrinsic activity or toxicity of the material,” Tanguay said.
Researchers have also exposed larval zebrafish to nanomaterials to assess interactions with more developmentally advanced bodily structures. “The larva is a very interesting stage of the fish” said Sijie Lin, a nanomedicine and nanotoxicology researcher at Tongji University. “They’ve hatched from their chorion; they can swim freely; they start to eat food that surrounds them; and they start to interact with the water environment more and more. So in this case, there are a few so-called biological barriers that are very important.”
To evaluate the effects of nanomaterials on one of these barriers, the skin, Lin’s team has taken advantage of the larva’s transparency. They used transgenic zebrafish larvae featuring macrophages and neutrophils labeled with green fluorescent protein (GFP), staining them bright green in an otherwise colorless sea of cells. “Whenever the macrophages and neutrophils start moving, that's an indication that there's something wrong,” Lin said. “Under the [fluorescent] microscope, you will see all the immune cells trying to concentrate in the area and trying to do damage control.”
The researchers exposed the larvae to a collection of cobalt oxide-based nanoparticles, which gathered along the surface of the skin. They observed that certain particle designs that form reactive oxygen species (ROS) showed both direct skin damage and a rush of macrophages and neutrophils to the injury site (9).
Transgenic zebrafish lines are hugely valuable in visualizing the biological processes underpinning toxicity. “People will say, well, the fish died. But that means there's a lot of things that already occurred before, so you missed the chance to really track anything down in real time,” Lin said. To probe the mechanism further, the researchers measured changes in gene expression in the zebrafish upon exposure to the toxic nanoparticles. They observed that genes involved in inflammation were upregulated, highlighting effects on a pathway that is highly conserved across species.
Interestingly, as zebrafish larvae rely on their skin to take up oxygen before their gills have fully formed, this structure is considered a suitable analogue of the lung epithelium (10). “If you compare the zebrafish skin and the lung epithelium, the cell types and the overall tissue structure are very similar,” Lin said. Understanding the toxic effects of ROS-generating nanoparticles toward larval skin therefore provides cautious insight into their implications for human health. “Based on the result, there is so much other [work] that needs to be done for its translation value, but it's a starting point. It's a very important starting point,” Lin said.
Lin’s team has also evaluated the effects of nanomaterials on the zebrafish larva’s gastrointestinal (GI) tract, which is anatomically similar to the mammalian version. Researchers have developed nanomaterials that can be orally administered to image the GI tract, protect fragile proteins from the harsh conditions of the stomach, and shuttle drugs across the intestinal epithelium, motivating efforts to understand these nanomaterials’ potential for adverse interactions (11,12). “The GI tract has been proven to be such an important [system] for humans, not only just for GI tract health, but because of the gut-brain axis, the gut-liver axis, and all these connections among different organs,” Lin said. “With any type of oral drug we introduce, we need to be very careful.”
People will say, well, the fish died. But that means there's a lot of things that already occurred before, so you missed the chance to really track anything down in real time.
- Sijie Lin, Tongji University
Zebrafish larvae make for a straightforward oral exposure model, indiscriminately consuming substances suspended in the water. By administering cerium oxide nanomaterials of different shapes in this way, Lin and colleagues observed that larvae that ingested long and pointy nanomaterials showed stunted growth and development, while shorter and rounder particles did not produce this effect (13). When they analyzed the fish under the microscope, they saw that the needle-like nanorods had pierced through the cells lining the GI tract. “The key mechanism of injury is actually physical,” Lin said. Similarly shaped “long aspect ratio” nanomaterials have been explored for oral drug delivery (14-16).
Lin’s team also found that ingestion of graphene-based nanomaterials of different sizes had distinct effects on the composition of the zebrafish gut microbiome (17). It isn’t yet clear if the nanomaterial directly kills a specific species of microbe, alters the physiological environment of the gut in a way that influences its microbial community, or acts through another mechanism, but “it’s something that’s worth digging into,” Lin said.
Benign by design
A nanomaterial’s size, shape, surface coating, and other parameters can be varied simultaneously to influence both its activity as a nanomedicine and its potential toxicity, yielding many possible design combinations. Fortunately, tiny zebrafish are perfect for screening large libraries of nanodrugs. “We’ve leaned heavily into the fact that they're small. Some people get upset, ‘Oh, they're so small.’ Well, that's why we invented microscopes,” Tanguay said. “We can scale our assays down to just tens of microliters in a small well and then do the exposures and measure all those endpoints in individual animals in four or five days. That's really valuable for folks who don't want to give us a half a million dollars-worth of test materials.”
To eliminate bottlenecks during high-throughput screening, the researchers have streamlined the entire experimental workflow. “We have these large tanks that house hundreds if not thousands of individual fish. And we're able to do these mass spawnings every day, where we get 10,000 to 100,000 embryos in these devices that we built,” Tanguay said. Her team has automated the removal of the protective chorion from the embryos. “Once you take the chorion off, they're very fragile. So, then we had to build robots that can carefully identify them, pick them up, and, one at a time, put them into multiwell plates,” Tanguay said (18). “And then we have other robots that automatically do the dosing at very high precision and speed.”
Lin also uses robotic technology to prepare his experiments and then relies on high-content imaging instruments, typically used by pharmaceutical companies to examine cells, to photograph the zebrafish embryos at a rate of approximately 1,000 images per hour. “The problem became how to analyze those images because now you have tens of thousands of images every week, and we cannot imagine people sitting there looking at images one by one in front of a computer,” Lin said.
To solve this problem, Lin developed artificial intelligence algorithms to automatically recognize and sort the images based on the morphology of the zebrafish (19). Similarly, Tanguay uses cameras that track motion and software that analyzes the distance traveled in response to the stimulus to automatically measure motor activity in thousands of zebrafish per day (20).
The results of these experiments can not only pinpoint the concerning nanomaterials in a group, but help to uncover structure-toxicity relationships. “High-throughput screening allows us to make sense of the complex interactions between these physicochemical properties,” Wilhelm said. “The hope is that we can use this in the end for engineering safer and more effective materials, so that we basically have a database now that allows us to identify trends that then can be used to inform the design and engineering of next-generation nanomedicines.”
Researchers have termed this approach “safe by design” or “benign by design” (21). “Just telling us something's dangerous, I think that's not enough,” Lin said. “You need to have a clear connection between certain features of the nanoparticle and its consequences for toxicity first. By establishing this, then you can think about how to change that property and to validate whether the toxicity also decreases.” For example, Lin and his colleagues observed that zinc oxide and copper oxide nanoparticles release toxic metal ions when they dissolve and found that doping in small amounts of iron reduced the nanoparticles’ dissolution and detrimental effects on zebrafish embryos (22,23).
Tanguay’s team is working to develop machine learning models that can learn from the screening data which features are associated with toxicity in order to predict whether or not a nanomaterial design is toxic. In one study, they identified concentration, shape, surface area, and aggregation as the variables most predictive of a nanomaterial’s toxicity (24). “We need to nail down the bare minimum parameters that everyone should have, before they even make a material, to go through these models and say, okay, these will cause cytotoxicity, or these will kill the zebrafish or potentially cause behavior effects,” Truong said.
Tanguay hopes that this work will enable the proactive development of safer nanomaterials. “If we can predict the outcomes using a sensitive biosensor, and we gain confidence that it's sensitive to all kinds of modes of action, then maybe we can use that to inform the way we synthesize these materials,” Tanguay said. “Maybe at some point, we won't have to test anything. That would be the ultimate goal.”
Lin aims to shed light on how nanotoxicology can be used to improve nanomedicine, and in doing so, continue to create harmony between two fields that at times have been considered to be in opposition (25). “One is to eliminate anything that's toxic; another is to create something that will help us fight diseases,” he said. “When you start to think about it, we are actually sharing the same purpose, which is trying to create a better world.”
- Tanguay, R.L. The rise of zebrafish as a model for toxicology. Toxicol Sci 163, 3-4 (2018).
- Weissig, V., Elbayoumi, T., Flühmann, B., & Barton, A. The growing field of nanomedicine and its relevance to pharmacy curricula. Am J Pharm Educ 85, 8331 (2021).
- Hou, X., Zaks, T., Langer, R., & Dong, Y. Lipid nanoparticles for mRNA delivery. Nat Rev Mater 6, 1078-1094 (2021).
- Lopes, T.S., Alves, G.G., Pereira, M.R., Granjeiro, G.M., & Leite, P.E.C. Advances and potential application of gold nanoparticles in nanomedicine. J Cell Biochem 120, 16370-16378 (2019).
- Truong, L. et al. Residual weakly bound ligands influence biological compatibility of mixed ligand shell, thiol-stabilized gold nanoparticles. Environ Sci: Nano 4, 1634-1646 (2017).
- Truong, L. et al. Systematic determination of the relationship between nanoparticle core diameter and toxicity for a series of structurally analogous gold nanoparticles in zebrafish. Nanotoxicology 13, 879-893 (2019).
- Zhao, X. et al. Functionalized graphene oxide nanoparticles for cancer cell-specific delivery of antitumor drug. Bioconjug Chem 26, 128-136 (2015).
- Lopez, R.M., White, J.R., Troung, L., & Tanguay, R.L. Size- and oxidation-dependent toxicity of graphene oxide nanomaterials in embryonic zebrafish. Nanomaterials 12, 1050 (2022).
- Peng, G. et al. Redox activity and nano-bio interactions determine the skin injury potential of Co3O4?based metal oxide nanoparticles toward zebrafish. ACS Nano 14, 4166-4177 (2020).
- Rong, J., He, Y., Tang, J., Qiao, R., & Lin, S. “Fishing” nano-bio interactions at the key biological barriers. Nanoscale 13, 5954-5964 (2021).
- Asad, S., Jacobsen, A.C., & Teleki, A. Inorganic nanoparticles for oral drug delivery: opportunities, barriers, and future perspectives. Curr Opin Chem Eng 38, 100869 (2022).
- Date, A.A., Hanes, J., & Ensign, L.M. Nanoparticles for oral delivery: design, evaluation and state-of-the-art. J Control Release 240, 504-526 (2016).
- Lin, S. et al. Aspect ratio plays a role in the hazard potential of CeO2 nanoparticles in mouse lung and zebrafish gastrointestinal tract. ACS Nano 8, 4450-4464 (2014).
- Karimi, M. et al. Carbon nanotubes part II: a remarkable carrier for drug and gene delivery. Expert Opin Drug Deliv 12, 1089-1105 (2015).
- Uskoković, V., Lee, K., Lee, P.P., Fischer, K.E., & Desai, T.A. Shape effect in the design of nanowire-coated microparticles as transepithelial drug delivery devices. ACS Nano 6, 7832-7841 (2012).
- Liu, W. et al. Rod-like mesoporous silica nanoparticles facilitate oral drug delivery via enhanced permeability and retention effect in mucus. Nano Res 15, 9243-9252 (2022).
- Lu, K. et al. Biological uptake, distribution, and depuration of radio-labeled graphene in adult zebrafish: effects of graphene size and natural organic matter. ACS Nano 11, 2872-2885 (2017).
- Mandrell, D. et al. Automated zebrafish chorion removal and single embryo placement: optimizing throughput of zebrafish developmental toxicity screens. SLAS Technol 17, 66-74 (2012).
- Dong, G. et al. Deep learning-enabled morphometric analysis for toxicity screening using zebrafish larvae. Environ Sci Technol (2023). In press.
- Shen, Q. et al. Rapid well-plate assays for motor and social behaviors in larval zebrafish. Behav Brain Res 391, 112625 (2020).
- Schwarz-Plaschg, C., Kallhoff, A., & Eisenberger, I. Making nanomaterials safer by design? NanoEthics 11, 277-281 (2017).
- Xia, T. et al. Decreased dissolution of ZnO by iron doping yields nanoparticles with reduced toxicity in the rodent lung and zebrafish embryos. ACS Nano 5, 1223-1235 (2011).
- Naatz, H. et al. Safe-by-design of CuO nanoparticles via Fe-doping, Cu-O bond lengths variation, and biological assessment in cells and zebrafish embryos. ACS Nano 11, 501-515 (2017).
- To, K.T., Truong, L., Edwards, S., Tanguay, R.L. & Reif, D.M. Multivariate modeling of engineered nanomaterial features associated with developmental toxicity. NanoImpact 16, 100185 (2019).
- Bondarenko, O. et al. Nanotoxicology and nanomedicine: the yin and yang of nano-bio interactions for the new decade. Nano Today 38, 101184 (2021).