Study Of Quantitative Structure-activity Relationship Models For Marine Acticancer Agents

For a long time, human have committed to the study of anticancer drugs. Marine is more and more recognized as a source of extremely potent natural anticancer agents due to its unique ecosystem, biological diversity and molecular structure novelty of the compounds. The development of anticancer agents derived from marine life is quick and marine life has become important direction. This paper had built QSAR models of marine Fascaplysin analogues, Hydroximinosteroid analogues and Ascidiathiazone nanlogues to clarify the molecular mechanism using MEDV-13、indicator descriptor and ESCI extract structural information and subsequently using BSR、GA and Stepwise Regression Analysis to select descriptor, which can provide important theoretical basis for design and development of antitumor agents. Therefore, we focus on the above three nanlogues:(1) Fascaplysin drived from the sponge, an indole alkaloid, was found to be a specific cyclin-dependent kinase 4 (CDK4) inhibitor, although it was able to inhibit the proliferation of tumor cells, its toxicity limited its clinical trials. In the third chapter, the Molecular Electronegativity Distance Vector (MEDV-13) used to describe the chemical structure of 44 Fascaplysin analogues, quantitative structure - property relationship (QSAR) model of six parameters on IC50 for CDk4 was built. LOO method and external test set were used to validate the model. The results showed that the model had good estimation ability and predictive power for the external samples. The present study also indicated that groups such as =C-(-C-)、>C-、>N-、-O- and the position of the benzene ring on biphenyl were main affect on IC50 and para- benzene approved IC50. We further analyze the fraction contribution of descriptors. The results indicated that although x15 (interaction of -C-、=C- and >C- ) had a main affect to IC50, other descriptors with low fraction contribution could not be neglected. (2) Rencently, many unique steroid compounds were found in the sponge and some showed potent cytotoxicity to cancer cells by test. In the fouth chapter, a new ESCI index was used to investigate quantitative relationship between molecular structure of 17 marine steroid analogues and corresponding activity IC50. The obtained QSAR model demonstrated satisfactory predictability with r=0.9546 and q=0.9244. Analysising the molecular structures of these type, we knew the molecular size and oxime (=N-OH) have significant impact on the molecular activity.Ascidiathiazones, thiazine-containing 2-quinolinequinone carboxylic acid derivatives isolated from a New Zealand ascidian Aplidium sp, have anti-proliferative activity. 2D and 3D quantitative structure- activity relationship analysis (QSAR) studies were applied on a set of quinolinequinones to model and understand their anti-proliferative activity against tumor cell HL60. CoMFA model had five principal components with q~2, r~2, steric field and electrostatic field being 0.456, 0.891, 0.567and 0.433, respectively. Steric contour and electrostatic contour interpreted vividly the relationship between structure and activity. The low LOO cross-validated q~2 (0.456) prompted us to enlarge this study at 2D level. The 2D-QSAR model had good predictive ability and stability with r~2=0.8929 and q~2=0.8332. The results showed that 2D-QSAR and 3D-QSAR methods can complement each other to the analysis of compounds.