In vitro characterization of toxicity of complex pesticide mixtures in support of contaminated site risk assessment
Date
2024-01-24
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
ORCID
Type
Thesis
Degree Level
Doctoral
Abstract
The continuous development and production of pesticides to meet agricultural demands has left thousands of contaminated sites that release significant concentrations of hundreds of pesticides and their byproducts into the environment that may pose significant risks to human health. Simultaneous exposure to various pesticides has been confirmed by detecting a considerable level of several pesticides and their metabolites in the general population. However, the current chemical hazard assessment process that relies heavily on the extensive use of live animals (time-consuming, costly, and presents significant ethical concerns) is failing to provide toxicology information for the vast and ever-increasing numbers of pesticides and their mixtures to inform their risk assessment in a timely manner. Thus, there is an urgent need to develop new approach methodologies (NAMs) to replace, reduce, and refine (3Rs) live animal testing with high throughput and low-cost alternative approaches to improve hazard assessment. The main objectives of this thesis were to develop and evaluate a NAM system based on in vitro cell assays to characterize the toxicity of pesticides and their mixtures to inform mixture risk assessment.
First, a meta-analysis was conducted to identify the most appropriate in vitro model and endpoints to investigate the toxicity of pesticides and their mixtures. Across 66 studies, 108 pesticides were assessed on ten human-derived cell types at four endpoints. The correlation between all the available half-maximal inhibitory concentrations (IC50s) for pesticides presented in the scientific literature and in vivo risk guidelines acceptable daily intake (ADI) and no observed effect level (NOEL) or no observed adverse effect level (NOAEL) was investigated using the Spearman correlation analysis and linear regression models. Amongst the ten human cell lines, human neuroblastoma cells (SH-SY5Y) were the most sensitive and consistent in results in evaluating the toxicity of pesticides. Only the IC50s derived from SH-SY5Y cells, using MTT-24 & 48 h (MTT: (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay- the most used assay) correlated (rho = 0.56 - 0.79; p <0.05) with in vivo guidelines, ADIs, and NOEL/NOAELs for pesticide toxicity. This chapter's key finding was that SH-SY5Y cells could be an appropriate cell model for the characterization of the relative risk of pesticides or their mixtures in vitro.
Following the investigation of the most proper cell model, the toxicity of four banned organochlorine pesticides (OCPs) (aldrin, dieldrin, heptachlor, and lindane) and three registered herbicides (trifluralin, triallate, and clopyralid) detected in soil and groundwater of a legacy contaminated pesticide manufacturing and packaging site were evaluated using SH-SY5Y cells. Cell viability, lactate dehydrogenase (LDH) release, production of reactive oxygen species (ROS), and caspase 3/7 activity were evaluated following 24 h of exposure to six concentrations of these pesticides. In addition, differential gene expression was assessed using next-generation sequencing (RNA-Seq) at sublethal concentrations. Aldrin and heptachlor were the most toxic pesticides in SH-SY5Y cells. Dieldrin, lindane, trifluralin, and triallate exhibited moderate toxicity, and clopyralid was not toxic to SH-SY5Y cells. Aldrin and heptachlor induced their toxicity through damage to the cell membrane, while the toxicity of dieldrin was partially attributed to necrosis and apoptosis. Necrosis and apoptosis were not the mechanisms of toxicity for lindane, trifluralin, and triallate; however, the toxic effects of these compounds, at least in part, were involved in ROS generation. Gene expression profiles at the molecular level suggested decreased cell viability can be related to the inhibited proliferation and other mechanisms, such as mitophagy induced by moderate toxic insults. The anti-apoptotic properties of the most dysregulated genes confirmed the lack of caspase activation following exposure to most of the tested pesticides. In addition, enriched terms identified from the dysregulated genes were involved in pathways associated with the pathogenesis of neurodegenerative diseases. These findings align with in vivo experiments and epidemiological studies that confirmed the neurodegenerative effects of OCPs in animals and humans.
In the fourth chapter, the risk of mixtures of pesticides detected at the above legacy contaminated site was assessed using the binary weight-of-evidence (BINWOE) approach. Due to the lack of in vivo interaction data, the potential interactions between pesticides in 15 binary mixtures were investigated using the MTT assay and SH-SY5Y cells. Our findings showed that 60% of the binary mixtures elicited synergism, 27% displayed antagonism, and 13% showed additive effects in SH-SY5Y cells. Estimation of health risk using in vitro interactions data indicated that adults in industrial land use (current land use), and toddlers in commercial land use (future land use), were at the highest risk of inhaling the mixture of volatile pesticides (the most active exposure scenario at the contaminated site) compared to the other subpopulations. Incorporating interaction data into the risk assessment increased predicted risk by up to 20% and decreased the risk by 2% in some mixtures. Therefore, mixtures of organochlorine pesticides with the same mechanism of action did not always follow dose additivity on SH-SY5Y cells.
In conclusion, the positive relationship between toxicity data derived from SH-SY5Y cells, in vivo data, and safety guidelines, the neurotoxic nature of pesticides in animals and humans, and the neuron-like properties of SH-SY5Y cells served as foundations for using this cell line in the health risk assessment of neurotoxic pesticides. The cytotoxicity assessment and RNA-Seq results confirmed that SH-SY5Y could generate pesticide toxicity information relevant to humans. The potential interaction between pesticides, even at low concentrations approved by cytotoxicity and RNA-Seq tests, necessitated incorporating the interaction data in the risk assessment. Therefore, SH-SY5Y cells that produce biologically relevant data with low cost, high throughput, and no ethical concerns hold promise to replace and reduce the use of live animals in hazard assessment of numerous neurotoxic pesticides and their mixtures. Moreover, using SH-SY5Y cells in the hazard characterization of mixtures and the BINWOE approach can speed up the risk assessment of numerous mixtures whose actual risk is not adequately characterized for human health.
Description
Keywords
Pesticides, mixture, in vitro, cell lines, toxicity, risk assessment
Citation
Degree
Doctor of Philosophy (Ph.D.)
Department
Toxicology Centre
Program
Toxicology