Anticancer Active Components Identification From Dysosma Versipellis-Liquorice Based On Composition-activity Relationship

The extract of Dysosma versipellis(Bajiaolian in Chinese)and liquorice(Gancao in Chinese)mixture was researched in this work.The composition-activity relationship was studied based on chemonetrics,and the optimal quantitative composition-activity relationship(QCAR)model was combined with mean impact value(MIV)method to recognize the main anticancer components from the complex extract,providing reliable guidence for the study of bioactive compounds and new drug research and development based on Bajiaolian-Gancao mixture.The main work is as follow:1.A HPLC-UV analytical method was constructed on Waters e2695-2489 system to get the fingerprint spectrum of the extract,and this method was used to analylize 46 batches of extracts of Bajiaolian-Gancao mixture from different sources in China,and then,45 chromatogram peaks were designated as common peaks of all samples.These common components were measured by relative peak area with vanillin as internal standard,and the results showed that there existed remarkable differences in the contents of different components and different samples.2.The extract of Bajiaolian-Gancao mixture was qualitatively analyzed on UPLC-ESI-Q-TOF tandom mass spectrometer in negative mode,and according to the parent ion and fragment ion in MS and MS/MS spectral respectively and compared with lots of literature,except peak 15,the rest of 44 common peaks had been tentatively determined,among which 8 components were from Bajiaolian including lignans and flavonoids,and 36 components belonged to Gancao including triterpenoid saponins,flavonoids and coumarins.3.MTT(3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide)assay was employed to test the anticancer efficacy of the extract of Bajiaolian-Gancao mixture on human cervical carcinoma HeLa cells in vitro.The inhibition rate of the 46 samples varied from 56.18 to 93.35%,and at the same dose,most samples had good inhibitory effect against HeLa cells,and meanwhile,different samples yielded very discrepant inhibition rate.4.Back propagation neutral network and support veator regression algotithm were used to bulid QCAR models for Bajiaolian-Gancao mixture using the acquired chemical composition data as independent variable and the corresponding medicinal efficacy data as dependent variable,and genetic algorithm and particle swarm optimization algorithm were applied for parameter optimization.Considering fitting precision and generalization ability,the PSO-SVR model with g=0.1,C=5.3072 had the best performance which had R=0.9964,RMSE=0.0044 in train set and R=0.9509,RMSE=0.0183 in test set,and it can accurately unscramble the complicated composition-activity relationship of the herb pair.5.Based on the optimal PSO-SVR model,MIV method was introduced to evaluate the contribution of the 45 common constituents to the anticancer effect of the extract.According to the ranking of absolute MIV,10 compounds were identified as main anticancer components,including isoquercitrin,kaempferol-3-O-glucopyranoside,4’-demethylpodophyllotoxin glucoside and podophyllotoxin from Bajiaolian and Macedonoside A,licoflavonol,glycycoumarin,gancaonin A,angustone A and isoangustone A from Gancao.Among these 10 components,8 compounds have been reported with anticancer activity or mechanism of action,and for the other 2 compounds,the results of this work provide credible reference for the study of their bioactivity in future.