Integration of read-across and artificial neural network-based QSAR models for predicting systemic toxicity: A case study for valproic acid

Abstract

We present a systematic, comprehensive and reproducible weight-of-evidence approach for predicting the no-observed-adverse-effect level (NOAEL) for systemic toxicity by using read-across and quantitative structure-activity relationship (QSAR) models to fill gaps in rat repeated-dose and developmental toxicity data. As a case study, we chose valproic acid, a developmental toxicant in humans and animals. High-quality in vivo oral rat repeated-dose and developmental toxicity data were available for five and nine analogues, respectively, and showed qualitative consistency, especially for developmental toxicity. Similarity between the target and analogues is readily defined computationally, and data uncertainties associated with the similarities in structural, physico-chemical and toxicological properties, including toxicophores, were low. Uncertainty associated with metabolic similarity is low-to-moderate, largely because the approach was limited to in silico prediction to enable systematic and objective data collection. Uncertainty associated with completeness of read-across was reduced by including in vitro and in silico metabolic data and expanding the experimental animal database. Taking the "worst-case" approach, the smallest NOAEL values among the analogs (i.e., 200 and 100 mg/kg/day for repeated-dose and developmental toxicity, respectively) were read-across to valproic acid. Our previous QSAR models predict repeated-dose NOAEL of 148 (males) and 228 (females) mg/kg/day, and developmental toxicity NOAEL of 390 mg/kg/day for valproic acid. Based on read-across and QSAR, the conservatively predicted NOAEL is 148 mg/kg/day for repeated-dose toxicity, and 100 mg/kg/day for developmental toxicity. Experimental values are 341 mg/kg/day and 100 mg/kg/day, respectively. The present approach appears promising for quantitative and qualitative in silico systemic toxicity prediction of untested chemicals.

Molecular Image-Based Prediction Models of Nuclear Receptor Agonists and Antagonists Using the DeepSnap-Deep Learning Approach with the Tox21 10K Library

“…Recent technological advances have focused on in silico approaches, such as quantitative structure-activity relationship (QSAR), based on the assumption that similar structures are associated with similar biological activities, taking advantage of their ability to accurately predict the toxicologically discrete values of the chemical or biological properties of molecules [5,[33][34][35][36][37]. However, the QSAR approach has the following disadvantages: (i) required skills and knowledge for feature extraction and selection, (ii) paucity of model interpretability, and (iii) low prediction performance due to the dependence on the choice of molecular descriptors and the prediction modeling algorithms [36,[38][39][40].…”

Green Toxicology: Connecting Green Chemistry and Modern Toxicology

“…These tools learn from large training sets of data and are evaluated and validated with an external data set . In one case study, researchers developed a systematic read-across protocol integrating artificial neural network (ANN)-based QSAR models to fill gaps in rat repeated-dose and developmental toxicity data . The integrated in silico approach was applied to the pharmaceutical valproic acid, a developmental toxicant in humans and animals.…”

“…50 In one case study, researchers developed a systematic read-across protocol integrating artificial neural network (ANN)-based QSAR models to fill gaps in rat repeated-dose and developmental toxicity data. 86 The integrated in silico approach was applied to the pharmaceutical valproic acid, a developmental toxicant in humans and animals. Structurally similar analogues to valproic acid were selected using the OECD QSAR toolbox.…”

“…Predicted NOAEL values were similar enough to experimental values to make this approach a promising novel non-testing approach to replace animal experimental studies. 86 Thus, such an integrated testing strategy can and should be applied early on in the product development process.…”

Comprehensive Analysis of Metabolic Changes in Male Mice Exposed to Sodium Valproate Based on GC-MS Analysis

“…Studies have shown that VPA can cause damage to the liver, 10–12 kidney, 13 heart, 14 and brain. 4 , 15 However, reports of VPA damage to other organs are rare, and the specific mechanism of VPA-induced organ toxicity has not been characterized. Therefore, the purpose and significance of this experiment is to comprehensively evaluate the toxic effect of VPA by studying the levels of metabolites in the intestine, lung, liver, hippocampus, cerebral cortex, inner ear, spleen, kidney, heart, and serum, and to preliminarily study the specific mechanism of organ toxicity induced by VPA.…”