Metabolic trajectory characterisation of xenobiotic-induced hepatotoxic lesions using statistical batch processing of NMR data

Abstract

Multivariate statistical batch processing (BP) analysis of ¹H NMR urine spectra was employed to establish time-dependent metabolic variations in animals treated with the model hepatotoxin, α-naphthylisothiocyanate (ANIT). ANIT (100 mg kg⁻¹) was administered orally to rats (n = 5) and urine samples were collected from dosed and matching control rats at time-points up to 168 h post-dose. Urine samples were measured via¹H NMR spectroscopy and partial least squares (PLS) based batch processing analysis was used to interpret the spectral data, treating each rat as an individual batch comprising a series of timed urine samples. A model defining the mean urine profile over the 7 day study period was established, together with model confidence limits (±3 standard deviation), for the control group. Samples obtained from ANIT treated animals were evaluated using the control model. Time-dependent deviations from the control model were evident in all ANIT treated animals consisting of glycosuria, bile aciduria, an initial decrease in taurine levels followed by taurinuria and a reduction of tricarboxylic acid cycle intermediate excretion. BP provided an efficient means of visualising the biochemical response to ANIT in terms of both inter-animal variation and net variation in metabolite excretion profiles. BP also allowed multivariate statistical limits for normality to be established and provided a template for defining the sequence of time-dependent metabolic consequences of toxicity in NMR based metabonomic studies.