An in silico approach for modelling T-helper polarizing iNKT cell agonists

Many analogues of the glycolipid alpha-galactosylceramide (α-GalCer) are known to activate iNKT cells through their interaction with CD1d-expressing antigen-presenting cells, inducing the release of Th1 and Th2 cytokines. Because of iNKT cell involvement and associated Th1/Th2 cytokine changes in a broad spectrum of human diseases, the design of iNKT cell ligands with selective Th1 and Th2 properties has been the subject of extensive research. This search for novel iNKT cell ligands requires refined structural insights. Here we will visualize the chemical space of 333 currently known iNKT cell activators, including several newly tested analogues, by more than 3000 chemical descriptors which were calculated for each individual analogue. To evaluate the immunological responses we analyzed five different cytokines in five different test-systems. We linked the chemical space to the immunological space using a system biology computational approach resulting in highly sensitive and specific predictive models. Moreover, these models correspond with the current insights of iNKT cell activation by α-GalCer analogues, explaining the Th1 and Th2 biased responses, downstream of iNKT cell activation. We anticipate that such models will be of great value for the future design of iNKT cell agonists.