Identification of individual structural fragments of N,N'-(bis-5-nitropyrimidyl)dispirotripiperazine derivatives for cytotoxicity and antiherpetic activity allows the prediction of new highly active compounds

Objectives: The objectives of this study were (i) to apply computer-based technologies to evaluate the structure of 48 NN'-(bis-5-nitropyrimidyl)dispirotripiperazines which belong to a new class of highly active antiviral compounds binding to cell surface heparan sulphates, (ii) to understand the chemical-biological interactions governing their activities, and (iii) to design new compounds with strong antiviral activity. Methods: The logarithm of 50% cytotoxic concentration (CC50) in GMK cells, of 50% inhibitory concentration IC50) against herpes simplex virus type 1, and of selectivity index (SI = CC50/IC50) was used to develop quantitative structure-activity relationships (QSARs) based on simplex representation of molecular structure. The QSAR model was applied to design new compounds. Two of these compounds were synthesized, physico-chemically characterized and tested for cytotoxicity and antiviral activity. Results: Statistic characteristics for partial least squares models allow the prediction of CC50, IC50 and SI values. The OSAR results demonstrate a high impact of individual structural fragments for antiviral activity. Molecular fragments that promote and interfere with antiviral activity were defined on the basis of the obtained models. Electrostatic factors (38%) and hydrophobicity (34%) were the most important determinants of antiherpetic activity. Using the established method, new potential dispirotripiperazine derivatives were computationally designed. Two of these computationally designed compounds were synthesized. The biological test results confirm the computationally predicted values of these compounds. Conclusions. The established QSAR model is suitable for the design of new antiherpetic compounds and prediction of their activity.