Innovative approaches for rapid antemortem diagnosis of prion disorders and to predict synergistic effects of drug combinations

The treatment of cancer is mostly based on a multidrug strategy. The administration of more than one drug at the same time aims to avoid the increase of drug resistant cells and reduce the dose of each drug of the combination, exploiting the synergistic effect. At the moment the evaluation of the synergistic effect is mainly performed using the Combination Index (CI) method and also Isobolograms (IB). The CI method is based on the assumption that the effect of the drugs is due only to the inhibition of enzyme kinetics (Chou et al. 1981). This assumption is not valid for any kind of drug in use; it is well known and studied that also other mechanisms can be involved. Cisplatin (CDDP) is one example of anticancer drug that has a different mechanism of action than direct effect on enzymes. Moreover, both CI and IB methods can only establish the effectiveness of the drug combination experimentally tested. We tried to overcome the limits of the classical approach using the Experimental Design (ED) and Artificial Neural Networks (ANNs); this approach, in fact, could not only allow to better quantify the degree of drug synergism of the tested combinations, but also to predict the effect of all possible drug doses, leading to the identification of the combination with the most powerful synergistic effect. An ANN is a powerful mathematical tool in the form of a computer software that recreates the structure and the functions of a biological brain. It is composed by logic units called “neurons” organized in 3 layers: input, hidden and output. After an appropriate “learning” process the ANN is able to predict the response for values of inputs never used before. The learning process is performed with a limited number of experiments that produce input data for the network called “training set”. A previous study (published by Journal of Inorganic Biochemistry) allowed us to evaluate anti-cancer activity of new copper(II)-phenantroline complexes (Pivetta et al. 2011) versus human hematologic (CCRF-CEM and CCRF-SB) and solid tumor-derived cell lines (K-MES-1 and DU-145); afterwards, ANN method was applied to binary mixtures of these new copper(II) complexes with CDDP tested against CCRFCEM. The latter study was performed thanks to the joined efforts with a group of chemists of the University of Cagliari and from the Masaryk University (Brno, Czech Republic). The results obtained confirmed our hypothesis and were recently published by Talanta (see following articles).