Mass Spectrum Selectivity-based Chromatogram Baseline-shift Elimination And Metabonomics Studies On Liver Toxicity From Traditional Chinese Medicine

For the LC-MS-based metabonomics research, gradient elution is often used andchromatogram baseline-shift is a common problem, which will have an influence onmultivariate analysis. Based on the high resolution and selectivity of mass spectrum,this paper attempted to propose a method to eliminate baseline-shift, i.e., mass spectrumselectivity-based chromatogram baseline-shift elimination (MSS-CBSE). The methodwas respectively investigated by simulated data, standard samples and actual urinesamples of rats. The results illustrated that the MSS-CBSE method was capable ofeliminating the chromatogram baseline, and the baseline obtained by MSS-CBSEmethod was true in the whole range of chromatogram, containing homoscedastic noisein non-signal ranges and heteroscedastic noise in chromatographic peaks. Baseline-shifthad a marked effect on the metabobomics model, displaying that it could lead weaklierdeviation between normal and treanment groups. Therefore, baseline-shift could lead toa reduction of the evaluated toxicity, and may result in wrong evaluation of drugtoxicity.Based on UPLC-ESI-MS-based metabonomics technology, we investigated theliver toxicity of Huang-yao-zi (HYZ). HYZ could disturb the normal metabolism of rats.The exogenous metabolite signals of HYZ increased with prolonged administration time,showing an accumulative effect. Meanwhile, the exogenous metabolites with lowmolecular weight decreased but the ones with high molecular weight increased and theconcentration of exogenous metabolites with high molecular weight rose notably,suggesting that the metabolic way of HYZ has changed. In metabonomics model ofHYZ, there was a clear separation of HYZ group from the normal one, indicating thatHYZ could cause liver toxicity of rats on the first day of administration, and livertoxicity was accumulated to the most serious on the third day.In metabomonics study of Chai-hu (CH), CH could cause the liver toxicity of ratseven at the lower dosage (25g/kg, twice a day), and the toxicity was strongest on the second day, but displaying a recovery tendency to the normal state on the third day.Moreover, the intensity of exogenous metabolites from CH changed with prolongedadministration time. The intensity of small molecule metabolites of CH had a negativecorrelation with the toxicity of CH, but the intensity of large ones had a positivecorrelation with the toxicity, hinting that the liver toxicity of CH was probably causedby the larger molecule components. The change of exogenous metabolites from CHsuggested that significant alterations occurred in the metabolic way of CH and the livertoxicity caused by low dosage of CH appeared a correlation with time and dosage.Finally, the small-scale expert system was generated based on the metabonomicsmodels of six standard hepatotoxicants, in order to evaluate and predict the liver toxicityfrom HYZ and CH. Based on the expert system, the prediction samples in HYZ groupwere most close to the ones of thioacetamide group, while the prediction samples of CHgroup weren’t close to any hepatotoxicant group. According to the classfication modelof thiacetamide, the degree for HYZ group deviating from thiacetamide and normalgroups became larger with prolonged administration day, indicating that the livertoxicity induced by HYZ was different from that induced by thiacetamide. Because theexpert system was apparently small, the toxicity prediction of Tradition ChineseMedicine (TCM) had some divations. Only further expand the range of expert system,can help the evaluation and prediction of the toxicity of TCM, and this need furtherstudy.