Use of computational toxicology tools to predict in vivo endpoints associated with Mode of Action and the endocannabinoid system: A case study with chlorpyrifos, chlorpyrifos-oxon and Δ9Tetrahydrocannabinol

Abstract

Currently, there is a lack of knowledge about the effects of co-exposures of cannabis, contaminated with pesticides like chlorpyrifos (CPF) and the toxic metabolite CPF-oxon (CPFO). CPF/CPFO residues, and ^Δ9Tetrahydrocannabinol (^Δ9THC), the main component in cannabis, are known to disrupt the endocannabinoid system (eCBS) resulting in neurodevelopmental defects. Although there are in vivo data characterizing CPF/CPFO and ^Δ9THC, there are mechanistic data gaps and deficiencies. In this study, an investigation of open access CompTox tools and ToxCast/Tox21 data was performed to determine targets relating to the modes of action (MOA) for these compounds and, given the available biological targets, predict points of departure (POD). The main findings were as follows: 1) In vivo PODs for each chemical were from open literature, 2) Concordance between ToxCast/Tox21 assay targets and known targets in the metabolic and eCBS pathways was evaluated, 3) Human Equivalent Administered Dose (EAD_Human) PODs showed the High throughput toxicokinetic (HTTK) 3 compartment model (3COMP) was more predictive of in vivo PODs than the PBTK model for CPF, CPFO and ^Δ9THC, 4) Age-adjusted 3COMP HTTK-Pop EAD_Human, with CPF and CPFO ToxCast/Tox21 AC₅₀ values as inputs were predictive for ages 0-4 when but not ^Δ9THC compared to in vivo PODs. 5) Age-related refinements for CPF/CPFO were primarily from ToxCast/Tox21 active hit-calls for nuclear receptors, CYP2B6 and AChE inhibition (CPFO only) associated with the metabolic pathway. Only one assay target (arylhydrocarbon hydroxylase receptor) was common between CPF/CPFO and ^Δ9THC. While computational refinements may select some sensitive events involved in the metabolic pathways; this is highly dependent on the cytotoxicity limits, availability of metabolic activity in the ToxCast/Tox21 assays and reliability of assay performance. Some uncertainties and data gaps for ^Δ9THC might be addressed with assays specific to the eCBS. For CPF, assays with appropriate metabolic activation could better represent the toxic pathway.

Keywords: Chlorpyrifos; Chlorpyrifos-oxon; HTTK-Pop; HTTK:, High throughput toxicokinetics; High throughput toxicokinetics; Pop, Population; ToxCast/Tox21; ToxCast/Tox21, CompTox Chemicals Dashboard High throughput Assays; Δ9Tetrahydrocannabinol.

Graphical abstract

Fig. 1

Natural logarithm fold-differences between 3COMP or PBTK models EAD_Human and in vivo PODs determined by allometric scaling (AS) or default 10-fold uncertainty factor (UF). The dotted lines describe the “perfect match” (ln(1) = 0) and “maximum uncertainty” (ln(10) = 2.3) due to intraspecies (human) variation between the in vivo- and httk model-based values. The vertical bars are the standard deviation of fold-differences.

Fig. 2

CPF and CPFO natural logarithm fold-differences between the EAD_Human with the 3COMP model and in vivo PODs with AS interspecies extrapolation. The dotted lines approximate the natural logarithm fold-difference between the in vivo AS-ELOEL/LOEL and the ToxCast/Tox21 assays that are not related to the MOA/eCBS. Assays below the lines for the respective chemicals are likely to be associated with the MOA/eCBS.