engleski

In silico prediction vs. Chromatographic determination of physicochemical properties of herbal therapeutics - statistical analysis

Evaluation of parameters such as lipophilicity, solubility and plasma protein binding through conventional approaches is time consuming and expensive, especially in high throughput screening. With the development of advanced prediction tools, in silico approaches enabled fast and inexpensive insight into characteristics of compounds based on their structure. Curcuminoids, andrographolides, boswellic acids and piperine, are part of food supplements used in patients suffering from inflammatory bowel diseases. The aim of this work was to evaluate the accuracy of in silico predicted log P, log S and plasma protein binding (PPB) values with the experimentally observed ones obtained using thin layer chromatography and high precision liquid chromatography with conventional reversed phase columns (C-18) and biomimetic columns with human serum albumin (HSA) and alpha-1-acid glycoprotein (AGP). These parameters are important to understand their pharmacokinetics which plays an important role in the therapy outcome. For comparison of experimentally observed and in silico calculated data for 10 examined compounds (curcumin, demethoxycurcumin, bisdemethoxycurcumin, α-boswellic acid, β- boswellic acid, 11-keto-β-boswellic acid , 3- acetyl-11-keto-β-boswellic acid, andrographolide, neoandrographolide and piperine), principal component analysis (PCA) and cluster analysis (CA) were used. Results of PCA and CA analysis showed that close relation between in silico observed values compared to experimentally observed ones, implying that in silico prediction can be used for fast insight into physicochemical properties of compounds.

Statistical analysis, biomimetic chromatography

nije evidentirano