Advances in PAH mixture toxicology enabled by zebrafish

Abstract

Polycyclic aromatic hydrocarbons (PAHs) are a class of organic compounds produced by a variety of petrogenic and pyrogenic sources. PAHs inherently occur in the environment in complex mixtures. The early life-stage zebrafish model is a valuable tool for high-throughput screening (HTS) for toxicity of complex chemical mixtures due to its rapid development, high fecundity, and superb sensitivity to chemical insult. Zebrafish are amenable to exposure to surrogate mixtures as well as extracts of environmental samples and effect-directed analysis. In addition to its utility to HTS, the zebrafish has proven an excellent model for assessing chemical modes of action and identifying molecular initiating and other key events in an Adverse Outcome Pathway framework. Traditional methods of assessing PAH mixture toxicity prioritize carcinogenic potential and lack consideration of non-carcinogenic modes of action, assuming a similar molecular initiating event for all PAHs. Recent work in zebrafish has made it clear that while PAHs belong to the same chemical class, their modes of action can be divergent. Future research should use zebrafish to better classify PAHs by their bioactivity and modes of action to better understand mixture hazards.

1. PAH mixture formation and transport in the environment

Polycyclic aromatic hydrocarbons (PAHs) are a class of compounds characterized by multiple fused benzene rings and may be substituted with functional groups. The composition of PAH mixtures in the environment can vary widely based on their source and the extent of weathering. Petrogenic PAH sources include petroleum and coal and are produced over a long period of time under moderate temperatures and intense pressure. Pyrogenic sources produce PAHs by incomplete combustion of carbon-based fuels, with forest fires, tobacco smoking, and high-heat cooking being major sources of human and environmental exposure [1, 2]. While the composition of petrogenic mixtures varies widely based on geologic origin and anthropogenic processing, petrogenic mixtures tend to be enriched in alkylated PAHs while pyrogenic mixtures have a higher proportion of unsubstituted PAHs [3]. Due to the large range of physical-chemical properties exhibited by PAHs, the composition of PAH mixtures in the environment changes over time as they are differentially affected by transport and degradation mechanisms through a process known as weathering. PAHs can be broken down by UV radiation, microbial degradation, or metabolism by larger flora and fauna with the resulting transformation products adding to the already diverse PAH chemical space [4, 5]. While most often used for forensic sourcing, these differences in mixture composition between sources and over time are directly relevant to hazard characterization of PAH mixtures in the environment. Use of fossil fuels and many other human activities have made PAHs ubiquitous environmental contaminants and ongoing research continues to expand our knowledge about the structural diversity of PAHs and their transformation products [6, 7].

Lighter molecular weight PAH structures (2–3 rings) are readily transported via the atmosphere, resulting in widespread global distribution. Moderate and large molecular weight PAHs may sorb to particulate matter, with inhalable particles 2.5 microns or smaller (PM_2.5) being a major source of atmospheric transport and human PAH exposure with location and seasonal variability [8]. PAHs can additionally make their way to aquatic and terrestrial ecosystems, often identified at particularly high levels in soil and sediments due to their hydrophobicity [9]. Movement into surface and ground water then allows further transport and the potential for additional accumulation into biota. All of these processes act in tandem to affect the composition of PAH mixtures in the environment, complicating characterization of their composition and providing opportunities for exposure through multiple routes.

2. Use of zebrafish in PAH mixture hazard assessment

2.1. Zebrafish as a model for high-throughput screening

Chemical hazard assessment relies on a variety of research models, animal and non-animal, with each presenting inherent advantages and disadvantages. Traditional mammalian models, while highly standardized and relevant to humans, tend to be cost-prohibitive, labor intensive, and may pose ethical concerns. The zebrafish (Danio rerio) has emerged as an ideal alternative model for high-throughput screening of chemicals for drug discovery, green chemistry, and for prioritizing chemicals for risk assessment due to its ease of use, translation to human health, and high sensitivity [10–14]. Zebrafish colonies are relatively easy to maintain and spawn with high fecundity, allowing for a large number of chemicals to be screened at a time. They also share high genetic similarity with humans and other vertebrates in addition to many of the same major organs including heart, liver, and pancreas [15, 16]. Since many PAHs are rapidly metabolized, often to reactive intermediates, the metabolic competency of developing zebrafish is a key strength when assessing the toxicity of PAHs and their mixtures. As a systems-based model, zebrafish have utility in hazard characterization of PAH mixtures with constituents which act via diverse mechanisms (Figure 1).

Figure 1.

Examples of key applications of the zebrafish model in toxicology of environmental chemicals and mixtures.

The sensitivity of zebrafish is particularly displayed during early development, with major organs beginning to develop just a few hours post-fertilization (hpf). Organogenesis in zebrafish is rapid, with major events like liver and intestine functionality by 72 and 96 hpf, respectively [17]. With many potential processes for a chemical to disrupt, abnormal development may present as physiological or behavioral defects [18–20]. Malformations may be visually observed in zebrafish throughout early development due to their optically clear tissues. Where a framework exists to characterize the hazard of mixtures, it is critical to have single chemical dose-response data as hazard references. The embryonic zebrafish model allows for prioritization of individual chemicals or groups of chemicals for further inquiry and its high sensitivity allows for assessment of differences in potency across structurally similar compounds [21–24].

2.2. Whole mixtures and effect-directed analysis

Strategies which directly combine the collection of environmental samples with bioassays are innovative and have proven useful for assessing hazard of real-world mixtures [25, 26]. Exposing an organism to a whole mixture from the environment is the ideal method to assess actual hazard as complete, quantitative analytical characterization of real-world exposures may be impossible. Typically, analysis of environmental samples will either quantify a set of known chemicals (targeted) or will compare chemicals identified in the sample to data from reference libraries to determine presence or absence of individual chemicals (non-targeted). As a non-targeted hazard identification technique, bioassay exposures to whole mixtures represent a unique complement to the available analytical chemistry techniques as the response of bioassays to chemicals in whole mixtures is not limited to chemicals identified a priori [27].

One area of research that has utilized whole mixture exposures extensively is oil spill science. Several studies have assessed the biological effects of crude oil in aquatic species local to the area of a spill and in the zebrafish surrogate model. The 2010 Deepwater Horizon oil spill was a major ecological event, releasing over 500 million liters of crude oil into the Gulf of Mexico [28]. Laboratory studies on effects of whole extract dilutions from this spill have been performed across several fish species, identifying significant multi-organ toxicities induced at exposure concentrations similar to those detected in the spill [29]. Cardiovascular toxicity and effects on metabolism and behavior were primary effects of exposure during early development [30–32]. Embryonic zebrafish exposed to extracts exhibited altered pronephros development, craniofacial malformations, and altered locomotor behavior across multiple studies [33]. The observed effects of exposure to crude oil mixtures, which contain a large PAH fraction, impact the local ocean ecology and commercial fisheries. In combination with techniques like toxicity identification evaluation and effect-directed analysis which seek to identify key toxicants, researchers have used zebrafish to improve our understanding of the mechanisms of exposure to spilled oil. By varying sample processing and exposure methodologies researchers have demonstrated the role of dissolved PAH mixtures in eliciting toxicity as opposed to direct contact to spilled oil droplets [34, 35]. By utilizing the zebrafish as a sensitive surrogate model, the hazard of exposure to crude oil can be rapidly assessed and used to inform remediation strategies following chemical releases [36].

While exposure to whole mixtures from the environment is a useful tool for hazard characterization, much can be learned from manipulating environmental samples prior to exposure as in the frameworks of toxicity identification evaluation and effect-directed analysis (EDA). By fractionating chemical mixtures, each sub-mixture can be screened to identify the bioactive fraction(s), simplifying the chemical space of interest. The process can be repeated until the number of chemical contributors is minimized, allowing identification and confirmation of the toxic compound(s).

The zebrafish model is particularly well suited to EDA studies for a variety of reasons. In addition to its sensitivity and high-throughput, only low-volumes are needed for exposure, reducing the quantity of environmental sample and labor required to generate samples and fractions. Another strength is the ability to measure both organism-level acute toxicity as well as mechanism-specific endpoints, facilitating identification of bioactive constituents [37]. The use of zebrafish in EDA has enabled considerable advancement in our understanding of PAH mixture toxicity. This method was employed to determine the developmentally toxic fraction of creosote contaminated porewater samples from the Elizabeth River estuary Superfund site in Virginia, USA [38]. Fractionation and subsequent exposures in zebrafish identified high concentrations of aryl hydrocarbon receptor (AHR) activating PAHs as the likely inducers of acute toxicity from porewater samples. Additionally, co-exposure with the weak Cyp1a inhibitor fluoranthene increased toxicity indicating synergistic effects between PAHs that activate AHR and those that inhibit the breakdown of ligands. The utility of zebrafish to capture complex interactions of PAH mixtures is a substantial advantage over in-vitro bioassays which are target- and thus endpoint-limited. Often the responsible toxicants identified by EDA in zebrafish are not routinely monitored or known to the researchers a priori. In one case, researchers used EDA and zebrafish to interrogate the toxicity of environmental samples from the Portland Harbor Superfund site in Portland, OR, USA. They found that despite being contaminants of concern at the site, PAHs were isolated in a non-toxic fraction, implicating contaminants other than PAHs in the toxicity of the whole mixture [39]. In a separate study investigating the degradation of PAH mixtures through a laboratory-scale remediation procedure, researchers used zebrafish and fractionation techniques to demonstrate that while the remedial technology indeed removed the target unsubstituted PAHs, oxygenated PAH derivates were formed which caused toxicity in the zebrafish model [40]. Together, these studies demonstrate the advantages of whole mixture and fractionation techniques. Traditional hazard characterization frameworks would fail to measure these toxicologically relevant contaminants. It should be noted that these approaches, while highly useful, rely on the method of sample collection, ability to fractionate compounds and quantify loss through sample processing, and an understanding of possible mixture interactions which could increase or decrease the perceived toxicity of a given fraction. Studies like these demonstrate the utility of zebrafish to improve our mechanistic understanding of both exposure and effects.

2.3. Surrogate mixtures and assessment of mixtures effects

In many cases, exposure to whole mixtures is infeasible and other approaches to hazard characterization must be used. One alternative technique applied by regulatory agencies in the United States of America and the European Union is the use of sufficiently similar or surrogate mixtures, defined by the USEPA as a “mixture that is very close in composition to the mixture of concern” [41, 42]. In a recent study, authors used zebrafish in combination with in-vitro cell assays to compare approaches used to form surrogate mixtures. They found that the best approach prioritized chemicals by synthesizing information about exposure and hazard potential [23]. Researchers at Oregon State University previously used the surrogate mixture approach to assess the toxicity of PAHs at the Portland Harbor Superfund site. A surrogate mixture of 10 PAHs was generated, consisting of those most abundant and commonly detected in surface waters within the Superfund site [21]. When tested individually, three constituents were bioactive in zebrafish, while the other seven were not. When tested in two sub-mixtures – one with active compounds, the other with inactives – both mixtures elicited developmental toxicity suggesting mixture interactions between compounds inactive through single chemical exposures. Studies like these highlight the value of considering not only the toxicity of individual compounds, but their combined effects when exposed at environmentally-relevant ratios.

Laboratory exposures to pairwise combinations of PAHs can also be used to deduce mixtures effects. Many, but not all, PAHs activate the AHR pathway inducing teratogenicity, tumorigenesis, and immunotoxicity with increased or decreased toxicity depending on how readily metabolized the parent compound is. Importantly, activation of the AHR pathway induces expression of several xenobiotic metabolizing enzymes, including cytochrome P450s. Cytochrome P450 1a (Cyp1a in zebrafish) is a key enzyme in the bioactivation of some PAHs. In these cases, a PAH both induces non-metabolizing downstream events of AHR activation and induces its own elimination or bioactivation, potentially causing toxicity from both events. PAHs like fluoranthene can inhibit Cyp1a, reducing its capacity to biotransform other co-exposed compounds, either reducing overall toxicity by preventing their bioactivation or increasing overall toxicity by reducing the elimination of toxic parent PAHs [43–45]. Understanding the role of biotransformation in the toxicity of a given PAH is important to better predict its toxicity in a complex mixture which may contain both AHR/CYP activators and inhibitors.

3. Approaches to predict PAH mixtures toxicity using zebrafish

3.1. Relative potency factor approach and divergent toxicity mechanisms

The classic approach for predicting risk of PAH mixtures utilizes relative potency factors (RPF) with benzo[a]pyrene (BaP) as the reference PAH and focuses on carcinogenicity as the primary endpoint. While carcinogenicity is an important toxicity endpoint in PAH chemical hazard assessment, approaches like RPF do not consider non-carcinogenic modes of action known to result from PAH exposure. The RPF approach assumes a common molecular initiating event – activation of the AHR. As mentioned previously, while many PAHs are known to activate the AHR, others do not. Additionally, due to ancestral genome duplication events, zebrafish have 3 AHRs – AHR1a, AHR1b, and AHR2, each of which may serve different non-overlapping roles[46]. The AHR2 isoform is most similar to that of mammals so, in the context of predicting human health outcomes, AHR2 is the most studied. Studies of PAH exposures in zebrafish lacking functional AHR2 have detected AHR2-independent toxicity and Cyp1a expression supporting the need for further classification of PAHs based on their level of AHR interaction, gross toxicity endpoints, and other known key events within an adverse outcome pathway (AOP) [47, 48].

In addition to acting through multiple mechanisms, PAH mixtures can exhibit dose dependent mixture interactions not easily predicted by single chemical dose response data [49]. These interactions can result from changes in gene expression and metabolism described in the preceding section, further complicating prediction of mixture toxicity through component-based models that assume additivity like the RPF approach.

3.2. Large-scale classification of PAHs to inform mixtures hazard prediction

Zebrafish are amenable to various molecular and visualization techniques to identify mechanisms that underlie PAH-induced toxicities. Assessment of spatiotemporal expression of key transcripts and proteins, reverse genetics tools, and use of -omics technologies have all proven fruitful for identifying molecular initiating events and key events within an adverse outcome pathway using the zebrafish model [48, 50]. A 2019 study by our group characterized the hazard of 123 PAHs based on zebrafish developmental toxicity screening, Cyp1a localization, and transcriptomic responses [51]. The compounds broadly grouped into two clusters: A and B. Cluster A consisted primarily of non-bioactive PAHs with few transcriptional responses while Cluster B contained PAHs with greater developmental toxicity, greater expression of cyp1a, and distinct AHR2-dependent Cyp1a protein expression in the skin. Additional studies have characterized similarity between zebrafish response to nitrated, oxygenated, and heterocyclic PAHs [52, 53]. Studies like these are highly useful for classification of PAHs at multiple levels of bioactivity to better understand their individual responses and inform mixtures assessments. With enough information on the observed toxicity and mechanisms of action of a broad suite of PAHs, the application of computational tools like quantitative structure-activity relationship models will enable scientists and regulators to predict the hazard potential of complex PAH mixtures.

4. The future of zebrafish in PAH mixture studies

Zebrafish have proven an ideal model for the assessment of PAH mixtures. As a surrogate model for mammals, zebrafish have been successfully used to provide rapid bioactivity data to inform remediation strategies. Molecular and visualization tools can be applied in zebrafish to identify and interrogate mechanisms of mixture toxicity. Future studies should use zebrafish to characterize the biological effects of as many individual PAHs as it is possible to obtain analytical standards for. The sooner this happens the sooner toxicologists will be able to predict the hazard potential of environmental PAH mixtures.