Concentration-dependent transcriptome of zebrafish embryo for environmental chemical assessment

Identification of Key Toxic Substances Considering Metabolic Activation: A Combination of Transcriptome and Nontarget Analysis

There have been numerous studies using effect-directed analysis (EDA) to identify key toxic substances present in source and drinking water, but none of these studies have considered the effects of metabolic activation. This study developed a comprehensive method including a pretreatment process based on an in vitro metabolic activation system, a comprehensive biological effect evaluation based on concentration-dependent transcriptome (CDT), and a chemical feature identification based on nontarget chemical analysis (NTA), to evaluate the changes in the toxic effects and differences in the chemical composition after metabolism. Models for matching metabolites and precursors as well as data-driven identification methods were further constructed to identify toxic metabolites and key toxic precursor substances in drinking water samples from the Yangtze River. After metabolism, the metabolic samples showed a general trend of reduced toxicity in terms of overall biological potency (mean: 3.2-fold). However, metabolic activation led to an increase in some types of toxic effects, including pathways such as excision repair, mismatch repair, protein processing in endoplasmic reticulum, nucleotide excision repair, and DNA replication. Meanwhile, metabolic samples showed a decrease (17.8%) in the number of peaks and average peak area after metabolism, while overall polarity, hydrophilicity, and average molecular weight increased slightly (10.3%). Based on the models for matching of metabolites and precursors and the data-driven identification methods, 32 chemicals were efficiently identified as key toxic substances as main contributors to explain the different transcriptome biological effects such as cellular component, development, and DNA damage related, including 15 industrial compounds, 7 PPCPs, 6 pesticides, and 4 natural products. This study avoids the process of structure elucidation of toxic metabolites and can trace them directly to the precursors based on MS spectra, providing a new idea for the identification of key toxic pollutants of metabolites.

A novel molecular pathway of lipid accumulation in human hepatocytes caused by PFOA and PFOS

Exposed to ubiquitously perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS) has been associated with non-alcoholic fatty liver disease (NAFLD), yet the underlying molecular mechanism remains elusive. The extrapolation of empirical studies correlating per- and polyfluoroalkyl substance (PFAS) exposure with NAFLD occurrence to real-life exposure was hindered by the limited availability of mechanistic data at environmentally relevant concentrations. Herein, a novel pathway mediating hepatocyte lipid accumulation by PFOA and PFOS at human-relevant dose (<10 μM) was identified by integrating CRISPR-Cas9 genome screening, concentration-dependent transcriptional assay in HepG2 cell and epidemiological data mining. 1) At genetic level, nudt7 showed the highest enriched potency among 569 NAFLD-related genes, and the transcription of nudt7 was significantly downregulated by PFOA and PFOS exposure (<7 μM). 2) At molecular pathway, upon exposure to ≤10-4 μM PFOA and PFOS, the downregulation of nudt7 transcriptional expression triggered the reduction of Ace-CoA hydrolase activity. 3) At cellular level, increased lipids were measured in HepG2 cells with PFOA and PFOS (<2 μM). Overall, we identified a novel mechanism mediated by transcriptional downregulation of nudt7 gene in hepatocellular lipid increase treated with PFOA and PFOS, which could potentially explain the NAFLD occurrence associated with exposure to PFASs in humans.

Insights into the synergistic toxicity mechanisms caused by nano- and microplastics with triclosan using a dose-dependent functional genomics approach in Saccharomyces cerevisiae

The emergence of polystyrene (PS) nano- and microplastics (NMPs) and triclosan (TCS) as environmental contaminants has raised concerns about their combined toxicities to organisms, but the complex toxicity arising from their interactions and the underlying molecular mechanisms remain obscure to us. In this study, we comprehensively detected the combined toxicity of PS-NMPs and TCS via the dose-dependent yeast functional genomics profiling. Firstly, our findings demonstrated that the combined exposure to PS-NMPs and TCS elicited a synergistic toxic effect in which the toxicity depended on the size of the PS-NMPs. Secondly, we found that TCS exposure, either alone or in combination with PS-NMPs, influenced lipid biosynthetic processes and ATP export pathways, while the unique responsive genes triggered by combined exposure to TCS and PS-NMPs are significantly enriched in mitochondrial translation, ribosomal small subunit assembly, and tRNA wobble uridine modification. Thirdly, our results demonstrated that point of departure (POD) at the pathway level was positively correlated with IC50, and POD was a more sensitive predictor of toxicity than the apical toxicity endpoints. More importantly, our findings suggested that the combined exposure of PS-NMPs in a size-dependent manner not only alleviated the harmful effects of TCS on glycerophospholipid metabolism, but also exacerbated its negative impact on oxidative phosphorylation. Collectively, our study not only provides new insights into the intricate molecular mechanisms that control the combined toxicity of PS-NMPs and TCS, but also confirms the effectiveness of the dose-dependent functional genomics approach in elucidating the molecular mechanisms of the combined toxicity of pollutants.

Empirical Characterization of False Discovery Rates of Differentially Expressed Genes and Transcriptomic Benchmark Concentrations in Zebrafish Embryos

High-throughput transcriptomics (HTTr) is increasingly applied to zebrafish embryos to survey the toxicological effects of environmental chemicals. Before the adoption of this approach in regulatory testing, it is essential to characterize background noise in order to guide experimental designs. We thus empirically quantified the HTTr false discovery rate (FDR) across different embryo pool sizes, sample sizes, and concentration groups for toxicology studies. We exposed zebrafish embryos to 0.1% dimethyl sulfoxide (DMSO) for 5 days. Pools of 1, 5, 10, and 20 embryos were created (n = 24 samples for each pool size). Samples were sequenced on the TempO-Seq platform and then randomly assigned to mock treatment groups before differentially expressed gene (DEG), pathway, and benchmark concentration (BMC) analyses. Given that all samples were treated with DMSO, any significant DEGs, pathways, or BMCs are false positives. As expected, we found decreasing FDRs for DEG and pathway analyses with increasing pool and sample sizes. Similarly, FDRs for BMC analyses decreased with increasing pool size and concentration groups, with more stringent BMC premodel filtering reducing BMC FDRs. Our study provides foundational data for determining appropriate experiment designs for regulatory toxicity testing with HTTr in zebrafish embryos.

Bioactivity assessment of organophosphate flame retardants via a dose-dependent yeast functional genomics approach

Organophosphate flame retardants (OPFRs) have been widely detected in multiple environment media and have many adverse effects with complex toxicity mechanisms. However, the early molecular responses to OPFRs have not been fully elucidated, thereby making it difficult to assess their risks accurately. In this work, we systematically explored the point of departure (POD) of biological pathways at genome-wide level perturbed by 14 OPFRs with three substituents (alkyl, halogen, and aryl) using a dose-dependent functional genomics approach in Saccharomyces cerevisiae at 24 h exposure. Firstly, our results demonstrated that the overall biological potency at gene level (PODDRG20) ranged from 0.013 to 35.079 μM for 14 OPFRs, especially the tributyl phosphate (TnBP) exhibited the strongest biological potency with the least PODDRG20. Secondly, we found that structural characteristics of carbon number and logKow were significantly negatively correlated with POD, and carbon number and logKow also significantly affected lipid metabolism associated processes. Thirdly, these early biological pathways of OPFRs toxification were found to be involved in lipid metabolism, oxidative stress, DNA damage, MAPK signaling pathway, and amino acid and carbohydrate metabolism, among which the lipid metabolism was the most sensitive molecular response perturbed by most OPFRs. More importantly, we identified one resistant mutant strain with knockout of ERG2 (YMR202W) gene participated in steroid biosynthesis pathway, which can serve as a key yeast strain of OPFRs toxification. Overall, our study demonstrated an effective platform for accurately assessing OPFRs risks and provided a basis for further green OPFRs development.

Identify organic contaminants of high-concern based on non-targeted toxicity testing and non-targeted LC-HRMS analysis in tap water and source water along the Yangtze River

Many organic pollutants were detected in tap water (TW) and source water (SW) along the Yangtze River. However, the potential toxic effects and the high-concern organics (HCOs) which drive the effect are still unknown. Here, a non-targeted toxicity testing method based on the concentration-dependent transcriptome and non-targeted LC-HRMS analysis combining tiered filtering were used to reveal the overall biological effects and chemical information. Subsequently, we developed a qualitative pathway-structure relationship (QPSR) model to effectively match the biological and chemical information and successfully identified HCOs in TW and SW along the Yangtze River by potential substructures of HCOs. Non-targeted toxicity testing found that the biological potency of both TW and SW was stronger in the downstream of the Yangtze River, and disruption of the endocrine system and cancer were the main drivers of the effect. In addition, non-targeted LC-HRMS analysis combined with retention time prediction results identified 3220 and 631 high-confidence compound structures in positive and negative ion modes, respectively. Then, QPSR model was further implied and identified a total of 103 HCOs, containing 35 industrial chemicals, 30 PPCPs, 26 pesticides, and 12 hormones in TW and SW, respectively. Among them, the neuroactive and hormonal compounds oxoamide, 8-iso-16-cyclohexyl-tetranor prostaglandin E2, E Keppra, and Tocris-0788 showed the highest frequency of detection, which were identified in more than 1/3 of the samples. The strategy of combining non-targeted toxicity testing and non-targeted LC-HRMS analysis will support comprehensive biological effect assessment, identification of HCOs, and risk control of mixtures.

Toxicity effects of pesticides based on zebrafish (Danio rerio) models: Advances and perspectives

Pesticides inevitably enter aquatic environments, posing potential risks to organisms. The common aquatic model organism, zebrafish (Danio rerio), are widely used to evaluate the toxicity of pesticides. In this review, we searched the Web of Science database for articles published between 2012 and 2022, using the keywords "pesticide", "zebrafish", and "toxicity", retrieving 618 publications. Furthermore, we described the main pathways by which pesticides enter aquatic environments and the fate of their residues in these environments. We systematically reviewed the toxicity effects of pesticides on zebrafish, including developmental toxicity, endocrine-disrupting effects, reproductive toxicity, neurotoxicity, immunotoxicity, and genotoxicity. Importantly, we summarized the latest research progress on the toxicity mechanism of pesticides to zebrafish based on omics technologies, including transcriptomics, metabolomics, and microbiomics. Finally, we discussed future research prospects, focusing on the combined exposure of multiple pollutants including pesticides, the risk of multigenerational exposure to pesticides, and the chronic toxicity of aquatic nanopesticides. This review provides essential data support for ecological risk assessments of pesticides in aquatic environments, and has implications for water management in the context of pesticide pollution.

Characterizing temporal variability and repeatability of dose-dependent functional genomics approach for evaluating triclosan toxification

Dose-dependent functional genomics approach has shown great advantage in identifying the molecular initiating event (MIE) of chemical toxification and yielding point of departure (POD) at genome-wide scale. However, POD variability and repeatability derived from experimental design (settings of dose, replicate number, and exposure time) has not been fully determined. In this work, we evaluated POD profiles perturbed by triclosan (TCS) using dose-dependent functional genomics approach in Saccharomyces cerevisiae at multiple time points (9 h, 24 h and 48 h). The full dataset (total 9 concentrations with 6 replicates per treatment) at 9 h was subsampled 484 times to generate subsets of 4 dose groups (Dose A - Dose D with varied concentration range and spacing) and 5 replicate numbers (2 reps - 6 reps). Firstly, given the accuracy of POD and the experimental cost, the POD profiles from 484 subsampled datasets demonstrated that the Dose C group (space narrow at high concentrations and wide dose range) with three replicates was best choice at both gene and pathway levels. Secondly, the variability of POD was found to be relatively robustness and stability across different experimental designs, but POD was more dependent on the dose range and interval than the number of replicates. Thirdly, MIE of TCS toxification was identified to be the glycerophospholipid metabolism pathway at all-time points, supporting the ability of our approach to accurately recognize MIE of chemical toxification at both short- and long-term exposure. Finally, we identified and validated 13 key mutant strains involved in MIE of TCS toxification, which could serve as biomarkers for TCS exposure. Taken together, our work evaluated the repeatability of dose-dependent functional genomics approach and the variability of POD and MIE of TCS toxification, which will benefit the experimental design for future dose-dependent functional genomics study.

Prediction of zebrafish embryonic developmental toxicity by integrating omics with adverse outcome pathway

New approach methodologies (NAMs), especially omics-based high-throughput bioassays have been developed rapidly, providing rich mechanistic information such as molecular initiation events (MIEs) and (sub)cellular key events (KEs) in adverse outcome pathways (AOPs). However, how to apply the knowledge of MIEs/KEs to predict adverse outcomes (AOs) induced by chemicals represents a new challenge for computational toxicology. Here, an integrated method named ScoreAOP was developed and evaluated to predict chemicals' developmental toxicity for zebrafish embryos by integrating four related AOPs and dose-dependent reduced zebrafish transcriptome (RZT). The rules of ScoreAOP included 1) sensitivity of responsive KEs demonstrated by point of departure of KEs (POD_KE), 2) evidence reliability and 3) distance between KEs and AOs. Moreover, eleven chemicals with different modes of action (MoAs) were tested to evaluate ScoreAOP. Results showed that eight of the eleven chemicals caused developmental toxicity at tested concentration in apical tests. All the tested chemicals' developmental defects were predicted using ScoreAOP, whereas eight out of the eleven chemicals predicted by ScoreMIE which was developed to score MIEs disturbed by chemicals based on in vitro bioassays data. Finally, in terms of mechanism explanation, ScoreAOP clustered chemicals with different MoAs while ScoreMIE failed, and ScoreAOP revealed the activation of aryl hydrocarbon receptor (AhR) plays a significant role in dysfunction of cardiovascular system, resulting in zebrafish developmental defects and mortality. In conclusion, ScoreAOP represents a promising approach to apply mechanism information obtained from omics to predict AOs induced by chemicals.

The path via pathway-based approaches towards safety assessment: A concise review

For decades, chemical safety assessment has been proposed to shift from animal testing to in vitro testing systems in response to the call for the 3R. In Europe, the answer was to combine various information sources in integrated testing strategies (ITS); In the US, it was in 2007 when the landmark report by the National Research Council put forward a vision of in vitro toxicity testing paradigm. Since then, efforts to develop pathway-based assessment framework have been on the track. In 2010, systems biology brought out a conceptual framework called adverse outcome pathway (AOP), which took one step further from toxicity pathway to regulatory toxicology. Computational modeling, high-throughput screening, high-content omics have all been approached to facilitate this progress. This paper briefly reviewed the achievement of pathway-based chemical assessment since 2007, discussed potential pitfalls and challenges that mechanism-driven chemical assessment may undergo, and presented future perspectives of safety assessment that is to be based on computational system biology.

High throughput data-based, toxicity pathway-oriented development of a quantitative adverse outcome pathway network linking AHR activation to lung damages

The quantitative adverse outcome pathway (qAOP) is proposed to inform dose-responses at multiple biological levels for the purpose of toxicity prediction. So far, qAOP models concerning human health are scarce. Previously, we proposed 5 key molecular pathways that led aryl hydrogen receptor (AHR) activation to lung damages. The present study assembled an AOP network based on the gene expression signatures of these toxicity pathways, and validated the network using publicly available high throughput data combined with machine learning models. In addition, the AOP network was quantitatively evaluated with omics approaches and bioassays, using 16HBE-CYP1A1 cells exposed to benzo(a)pyrene (BaP), a prototypical AHR activator. Benchmark dose (BMD) analysis of transcriptomics revealed that AHR gene held the lowest BMD value, whereas AHR pathway held the lowest point of departure (PoD) compared to the other 4 pathways. Targeted bioassays were further performed to quantitatively understand the cellular responses, including ROS generation, DNA damage, interleukin-6 production, and extracellular matrix increase marked by collagen expression. Eventually, response-response relationships were plotted using nonlinear model fitting. The present study developed a highly reliable AOP model concerning human health, and validated as well as quantitatively evaluated it, and such a method is likely to be adoptable for risk assessment.

Advantages of omics technology for evaluating cadmium toxicity in zebrafish

In the last decade, several advancements have been made in omics technologies and they have been applied extensively in diverse research areas. Especially in toxicological research, omics technology can efficiently and accurately generate relevant data on the molecular dynamics associated with adverse outcomes. Toxicomics is defined as the combination of toxicology and omics technologies and encompasses toxicogenomics, toxicoproteomics, and toxicometabolomics. This paper reviews the trend of applying omics technologies to evaluate cadmium (Cd) toxicity in zebrafish (D. rerio). Cd is a toxic heavy metal posing several environmental concerns; however, it is being used widely in everyday life. Zebrafish embryos and larvae are employed as standard models for many toxicity tests because they share 71.4% genetic homology with humans. This study summarizes the toxicity of Cd on the nerves, liver, heart, skeleton, etc. of zebrafish and introduces detailed omics techniques to understand the results of the toxicomic studies. Finally, the trend of toxicity evaluation in the zebrafish model of Cd based on omics technology is presented.

The phenotypic and transcriptomic effects of developmental exposure to nanomolar levels of estrone and bisphenol A in zebrafish

Estrone and BPA are two endocrine disrupting chemicals (EDCs) that are predicted to be less potent than estrogens such as 17β-estradiol and 17α-ethinylestradiol. Human exposure concentrations to estrone and BPA can be as low as nanomolar levels. However, very few toxicological studies have focused on the nanomolar-dose effects. Low level of EDCs can potentially cause non-monotonic responses. In addition, exposures at different developmental stages can lead to different health outcomes. To identify the nanomolar-dose effects of estrone and BPA, we used zebrafish modeling to study the phenotypic and transcriptomic responses after extended duration exposure from 0 to 5 days post-fertilization (dpf) and short-term exposure at days 4-5 post fertilization. We found that non-monotonic transcriptomic responses occurred after extended duration exposures at 1 nM of estrone or BPA. At this level, estrone also caused hypoactivity locomotive behavior in zebrafish. After both extended duration and short-term exposures, BPA led to more apparent phenotypic responses, i.e. skeletal abnormalities and locomotion changes, and more significant transcriptomic responses than estrone exposure. After short-term exposure, BPA at concentrations equal or above 100 nM affected locomotive behavior and changed the expression of both estrogenic and non-estrogenic genes that are linked to neurological diseases. These data provide gaps of mechanisms between neurological genes expression and associated phenotypic response due to estrone or BPA exposures. This study also provides insights for assessing the acceptable concentration of BPA and estrone in aquatic environments.

Evidence-based assessment on environmental mixture using a concentration-dependent transcriptomics approach

Development of new approach methodologies is urgently needed to characterize the likelihood that complex mixtures of chemicals affect water quality. Omics advances in ecotoxicology allow assessment on a broadest coverage of disrupted biological pathway by mixtures. Here the usefulness of transcriptomic analyses for evaluation of combined effects and identification of main effect components are explored. Two artificial mixtures (Mix 1 and Mix 2) were tested by a concentration-dependent reduced zebrafish transcriptome (CRZT) approach and toxicity bioassays using zebrafish embryos. Then, the toxicities and transcriptomic effects of 12 component chemicals on embryos were incorporated into additivity models to characterize the combined effects of chemicals in mixtures and to identify the main bioactive compounds. Mix 1 and Mix 2 displayed similar embryo toxicities (LD₅₀: 6.6 μM and 8.7 μM, respectively), however, Mix 2 elicited broader biological process perturbations and 5-fold higher transcriptome potency (point of departure eliciting a 20% pathway response, POD_path20) than Mix 1. The predicted mixture toxicities derived from additivity expectations deviated by 2-fold or less from the measured embryo toxicities except for the Jaw defect endpoint; most biological processes deviated by 3-fold or less. Finally, diclofenac (DFC) and propiconazole (PCZ) were identified as the main contributing components (≥80% explanation) to the embryo toxicity and biological process perturbations by Mix 1. While DFC and chlorophene (CLP) explained up to 80% of the embryo toxicities and biological effects of Mix 2 associated with development and Metabolism processes. The CRZT approach provides a powerful tool for assessment of biological pathways perturbed by chemicals in mixtures and for identification of main bioactive compounds.

Developmental toxicity and transcriptome analysis of 4-epianhydrotetracycline to zebrafish (Danio rerio) embryos

As a primary degradation by-product of tetracycline (TC), 4-Epianhydrotetracycline (4-EATC) has been detected frequently in the aquatic environment, which may pose a potential environmental risk to aquatic organisms. Up to now, however, the toxicology study on 4-EATC to aquatic organisms is limited. In the present study, in order to better understand the toxic mechanism of 4-EATC, developmental toxicity including lethal and sublethal effects of 4-EATC and TC were investigated. The results showed that the developmental toxicity of 4-EATC to zebrafish embryos was stronger than that of TC. The 96 h LC₅₀ value of 4-EATC to zebrafish embryos was 29.13 mg/L. Malformations seemed to be the most sensitive sublethal endpoint of 4-EATC exposure, and the 96 h EC₅₀ value was 8.57 mg/L. Transcriptome response of 4-EATC to zebrafish embryos was determined. The results showed that 430 different expression genes (DEGs) caused by 4-EATC, and most enriched in tryptophan (TRP) metabolism pathway. Annotation of DEGs in the TRP metabolism demonstrated that expression of 4 gene products in tryptophan metabolized along the kynurenine (KYN) pathway were changed. Disorder of TRP catabolism in KYN pathway was a potential mechanism of 4-EATC toxicity to zebrafish embryos.