Toxicogenomic Fingerprint Identification in Springtails to Assess Pesticide-Contaminated Soils

Abstract

Agricultural soils are often polluted with pesticide mixtures. In part, this pesticide pollution has led to a global decline in invertebrates, which provide ecosystem services essential to sustainable agriculture. An initial step in protecting non-target invertebrates is to determine the environmental risk of pesticide pollution to guide pesticide abatement efforts. Conventional environmental risk assessment measures the soil concentration of an extensive panel of pesticides using chemical analysis. This labor-intensive process does not indicate two essential aspects of the environmental risk of pesticide pollution, its bioavailable or hazardous fraction to non-target invertebrates. Bioanalytical tools can supplement conventional chemical screening to assess the environmental risk of pesticide pollution more accurately. A toxic substance, such as a pesticide, can alter the regulation of hundreds or thousands of genes. These gene regulation patterns are unique for a toxic substance and therefore known as its toxicogenomic fingerprint. Biomarkers are reference genes that can detect the presence of these toxicogenomic fingerprints. In toxicogenomic fingerprint-based pesticide monitoring, soil samples are sent to a testing facility where lab-reared animals are exposed to them. These sentinels can provide a read-out of their response to toxic exposure and function as a living probe to assess the bioavailable and hazardous fraction of the pollutant mixture. Folsomia candida (springtails) would be ideal for this role; it is easily reared in the lab, requires a small amount of soil compared to other model species, and has been a soil ecotoxicological model for decades. I choose to focus on toxicogenomic fingerprint development for neonicotinoid soil pollution as neonicotinoids are the most commonly used type of insecticides of the past three decades and among the most toxic class of pesticides to invertebrates. My research questions were divided into two categories: (1) How to identify toxicogenomic fingerprints? (2) Are biomarkers derived from toxicogenomic fingerprints robust indicators of neonicotinoid exposure under various stress conditions? In chapter 2, I investigated what would be the most optimal exposure time to obtain transcriptomic and proteomic (omic) data for toxicogenomic fingerprint identification in Folsomia candida. Moreover, I investigated if shifts in transcript and protein abundances from the same gene occurred simultaneously or were delayed and whether this delay would hamper our ability to identify toxicogenomic fingerprints from multi-omics data. I found that the 48-hour time point was the most opportune moment for toxicogenomic fingerprint identification. Moreover, the results indicated no time lag between shifts in gene transcript and protein levels relevant to the combined analysis of the omics data. In chapter 3, I sought to identify toxicogenomic fingerprints from the transcriptomic data obtained from Folsomia candida exposed to two binary mixtures of pesticides that were finely resolved for stress intensity. To this end, we developed a novel statistical framework to identify toxicogenomic fingerprints from mixture-exposure-transcriptomic data. This framework can is broadly applicable, such as in research on the effects of multiple drugs in human disease treatments. In chapters 4 and 5, I inhibited the presumed pesticide detoxification pathways in F. candida with metabolic inhibitors. In this manner, the metabolic inhibitors provided a stress test for the biomarkers found in the previous chapters. I demonstrated that multiple biomarkers are needed to assess pesticide exposure accurately. Lastly, in chapter 6, I discussed future perspectives for toxicogenomic fingerprint identification and application for the purpose of monitoring pesticide soil pollution.