Mass Spectrum-based Orthogonal Projection Method And Its Application In Metabonomics Of Traditional Chinese Medicine

The basic idea of metabonomics was to construct successful metabonomics modelsfor the evaluation and prediction of drug toxicity depending on disturbances ofendogenous metabolites in biofluids. However, there were inevitably a large amount ofdrug and its metabolites in the metabonomics samples (like urine), especially fortraditional Chinese medicine (TCM), which may have a bad impact on the accuracy androbustness of metabonomics model. In this paper, a novel mass spectrum-basedorthogonal projection (MSOP) method was proposed, aiming to eliminate the signals ofexogenous metabolites. The method was investigated from simulated data, standardsamples and urine samples of rats in response to Acetaminophen (APAP) treatments,respectively. Then based on this method and metabonomics technology, the livertoxicity of aqueous extract of Huang-yao-zi (HYZ) was studied.First, the detail of the MSOP theory and algorithm were proposed. In this section,the effect of prerequisite (orthogonality of MS signals) and constraint condition on thismethod were separately discussed using the simulated data. Then with the UPLC-MSdata of standard samples, the feasibility of this method was investigated. The resultsdemonstrated that no matter how complex the data sets were, the information ofexogenous metabolites could be effectively eliminated.Combining UPLC-ESI-MS-based metabonomics technology with the trantionaldrug toxicity assessment methods, we investigated the liver toxicity of APAP in rats andestablished the metabonomics model using principal component analysis (PCA). What’smore, we used the data of rat urine to further verify the feasibility of MSOP method ineliminating the exogenous metabolites. With the MSOP method, we finally foundseveral ion peaks of exogenous metabolites, including APAP、 APAP-sulfate、NAC-APAP、APAP-S-S-APAP、APAP-glucuronide and its isomers. From the PCA score plot, it could be seen that the urine samples of APAP group were completely separatedfrom the control group, indicating that obvious liver toxicity took place in the APAPgroup. This was consistent with the results of clinical biochemistry and liverhistopathological observation.Combining UPLC-ESI-MS-based metabonomics technology and MSOP methods,we investigated the liver toxicity of HYZ in rats. It was found from the score plot ofPCA before eliminating the exogenous metabolites that there was a distinct clusteringbetween the control and HYZ group and the longer the administration time was, thefarther the HYZ group was away from control group. What’s more, the urine samples ofDay1of HYZ group (HYZ1) were completely separated from the control group.However, from the score plot after eliminating the exogenous metabolites using theMSOP method, we could find that the samples of the control group and HYZ1overlapped with each other, which demonstrated that the exgenous metabolites had abad impact on the accuracy of metabonomics model. In addition, the urine samples ofHYZ2and HYZ3group were still far away from the control group though the distancewas closer, which indicated that liver injury of rats in the two groups apparentlyoccurred. This was consistent with the results of clinical biochemistry and liverhistopathological observation.All in all, the MSOP method proposed in this paper was effective and reliable toeliminate the interfering signals of exgenous metabolites and the metabonomics modelwas more interpretable and robust for the evaluation and prediction of drug toxicity.Therefore, with the MSOP method metabonomics technology will be of greatsignificance in the research of drug toxicology, especially in TCM.