The emergance of antioxidant defense systems is one of nature’s greatest accomplishments, since it was this that enabled the spreading of life throughout the oxygen rich environment, that is the surface of our planet. Antioxidants are defined as “any substance that, when present at low concentrations compared to those of an oxidisable substrate, significantly delays or prevents the oxidation of that substrate”. In the search for novel potential antioxidants on the basis of natural chemical scaffolds Habrant et al discovered that a pigment found in the edible mushrooms Xerocomus badius and Pisolithus tinctorus Norbadione A acts as a powerful antioxidant under the influence of Fenton reaction, UV and gamma radiation, and is also features an interesting novel scaffold, which differs from traditional natural antioxidants, such as flavonoids or polyphenols. With the purpose of establishing a clear relationship between the structure and activity of such compounds a series of pulvinic acid derivatives had been syntetized, and their antioxidant activity evaluated with the degradation of thymidine assay, caused by oxidative stress as a consequence of Fenton reaction, UV or gamma radiation. On the basis of these data they developed qunatitative structure-activity relationship (QSAR) models, which predicted also coumarines as potential antioxidants, so in the following work Wagner et al. synthesised and experimentally evaluated 79 new pulvinic acid derivatives, as well as 23 coumarine derivatives and 9 structurally less related standard antioxidants. Their obtained experimental data represent the basis for the presented doctoral thesis, in which we have developed robust and validated models for the prediction of antioxidant activity and transmembrane transport of derivatives of pulvinic acid and coumarine, and then employed these models in the search for novel antioxidants, capable of crossing the cell membrane. In the first study we have developed several QSAR models on the basis of the antioxidant dataset, using different modeling techniques that represent novel in silico tools, useful in the search for novel antioxidants. We have also developed and presented the first known method for the graphical evaluation of the applicability domain of QSAR models, developed using support vector regression. In the next study we have developed a pharmacophore model on the basis of selected active pulvinic acid derivatives, which represents the spatial arrangement of generalized chemical features, necessary for the antioxidant activity. Together with the previously developed QSAR models we have used the pharmacophore model for the virtual screening of compound libraries in search for novel antioxidants. We have selected 21 promissing compounds from commercial sources for the experimental evaluation of their antioxidant activity using four in vitro assays. 10 of the selected compounds displayed antioxidant activity in at least one of the assays; this does not only represent the discovery of novel antioxidants, but also proves the predictive power of the developed in silico models. In order to better define the antioxidant activities of novel compounds, the activities of a few standard antioxidants were experimentally evaluated as well; it was found that the activities displayed by the novel compounds were comparable to, and in some cases even greater than those of the standard antioxidans, or the novel compounds were active against a broader spectrum of oxidative species in comparison to the standard antioxidants. While the activity of compounds is an important aspect in drug design, it is not the only prerequisite for the actual biological activity, since bioavailability is also of great importance. This is the reason our third study was focused on the development of QSAR models for the prediction of the transport of compounds by the transmembrane transporter protein bilitranslocase from the superfamily of organic anionic transporters. We developed the models on the basis of experimental data on the bilitranslocase transport activity of various small organic compounds (anthocyanins, flavonols, nucleotides, nucleosides, nucleobases, various endogenous molecules, dyes and drugs) that were obtained in several previous studies. We then employed the developed models for the prediction of the transport activity of the 10 novel compounds that displayed antioxidant activity in our previous study; the models predicted all 10 compounds as active bilitranslocase substrates. In the context of the presented doctoral thesis we have developed novel in silico models for the prediction of the antioxidant activity of compounds, as well as the evaluation of their ability to cross the cellular membrane through bilitranslocase. Furthermore, we presented a novel method for the assessment of the applicability domain of QSAR models, developed using support vector regression. We successfully applied the developed in silico tools for the virtual screening of compounds, which resulted in the discovery of 10 novel antioxidants, whose activity we confirmed experimentally. Next to the novel in silico tools, methods and novel antioxidant compounds, the doctoral thesis also represents an example of the usefulness of in silico approaches in the screening and search for novel compounds with antioxidant properties (or rational drug design in general): on the basis of experimental data for a small set of compounds we developed several in silico tools, which has enabled us to rationally select the most promising novel antioxidants. Finally, we have also confirmed the quality of our selection by the experimental evaluation of the selected compounds.
|
