Metabolomic 1H NMR Analysis Of Mass Pyrrolizidine Alkaloid Poisoning Outbreak In Ethiopia, With Validation In The Mouse Model

An outbreak of several hundred cases of Hirmi Valley Liver Disease (HVLD), an often fatal toxic hepatitis, occurred in 2001-2011 in a cluster of rural villages in Tigray, Ethiopia. We have previously shown that it is principally caused by contamination of the food supply by the plant toxins pyrrolizidine alkaloids, including acetyllycopsamine (AL), and that exposure to DDT increases the susceptibility of residents to AL. Using 1H NMR metabolomics, we cross-sectionally analysed urine samples from HVLD cases and controls; controls stratified by exposure status; and samples collected from mice dosed with AL or AL and DDT in combination. 600 MHz 1H NMR spectra were acquired and multivariate analysis applied. Orthogonal-Partial Least Squares – Discriminant Analysis (PLS-DA) compared 10 case versus 7 control samples collected in 2008, 35 case versus 18 control samples collected in 2009 and 18 high-exposed versus 18 low-exposed control samples. PLS-DA models compared urinary spectra from C57BL/6J mice from acute (1x 1500 mg/kg AL) and chronic (8 x 500 mg/kg AL over 4 weeks) dosing experiments. Mice dosed with AL with or without pre-dosing p,p'-DDT were compared by univariate methods. Cross-validated models distinguished all compared sample classes. Discriminatory metabolites included tyrosine, bile acids, N-acetylglycoproteins, N-methylnicotinamide and formate, hippurate, p-cresol sulphate, p-hydroxybenzoate and 3-hydroxypropionic acid. Tyrosine and p-cresol sulphate were associated with both exposure and disease. Similar changes to tyrosine, one-carbon intermediaries and microbial associated metabolites were observed in the mouse model, with tyrosine correlated with the extent of liver damage. In this novel application of metabolomics to a mass poisoning episode we have identified intermediate markers of exposure and disease, some of which were also found in the dosing mouse model. Pathways identified here may be useful in translational research and as “exposome” signals.