Integrating in silico and in vitro approachers: Cytotoxicity and enzymatic activity of xenobiotics in different fish cell lines

Abstract

In the present study, five Quantitative Structure-Toxicity Relationship (QSTR) models were generated. Three QSTR models were developed for the cytotoxcity (pEC50) of diverse chemicals to Poeciliopsis lucida hepatocarcinoma cell line (PLHC-1) measured with three assays, namely ethoxyresorufin-O-deethylase enzyme (EROD), neutral red (NR) and 3-(4, 5-dimethylthiazol-2-yl)2, 5-diphenyl tetrazolium bromide (MTT) assays. EROD, NR and MTT assays reflects enzymatic activity, lysosomal and mitochondrial activity, respectively. The other two QSTR models were generated using the NR data for goldfish (GFS), Carassius auratus, scale tissue and fathead minnow (FHM), Pimephales promelas, epithelial cell line. Descriptors appearing in each model were selected via the tools implemented in QSARINS 2.2.1 software. All QSTR models were generated in line with the Organization of Economic Co-operation Development (OECD) principles and validated both internally and externally. pEC50,EROD[PLHC-1] model had one descriptor, pEC50,NR[GFS] had two and each of pEC50,NR[PLHC-1], pEC50,NR[FHM] and pEC50,MTT[PLHC-1] models had three descriptors from DRAGON. The external predictivity of the generated QSTR models were tested using structurally diverse chemicals with no experimental cytotoxicity data. Structural coverage of the pEC50,NR[PLHC-1], pEC50,NR[GFS], pEC50,MTT[PLHC-1], pEC50,EROD[PLHC-1] and pEC50,NR[FHM] models for the external set compounds were 98.6%, 95.0%, 98.3%, 92.1% and 89.1%, respectively. A moderate/strong correlation was observed between the experimental in vivo and predicted in vitro values for the external set chemicals, except fathead minnow. The generated QSTR models may provide an initial, rapid screening and prioritization of these diverse chemicals for the acute fish toxicity assessment and reduce the need for extensive in vivo toxicity testing.