The well-known diuretic amiloide displays antitumor and antimetastatic side-activities, which have been attributed to dual-inhibition of the urokinase plasminogen activator (uPA) and the human sodium hydrogen exchanger isoform 1 (hNHE1).
uPA plays a critical role in tumour cell invasion and migration and is a promising anti-metastasis drug target. 6-Substituted analogues of 5-N,N-(hexamethylene)amiloride (HMA) are potent uPA inhibitors that show anti-metastatic effects in vivo, while lacking the diuretic and anti-kaliuretic properties of amiloride. However, the compounds as a class display pronounced selectivity for human over mouse uPA, thus confounding interpretation of data from human xenografted mouse models of cancer. To understand the molecular basis of this selectivity molecular dynamics simulations and alchemical free energy perturbation calculations were performed using human and mouse uPA in complex with amiloride, HMA and 6-substituted analogues. The results were compared to the experimental enzyme inhibitory potencies.
At the outset of this study, the absence of a complete structure of hNHE1, prompted creation of an in silico model that could be used to design more potent amiloride based inhibitors. In this work, a homology model of hNHE1 was built, whose topology was very consistent with a recently solved first full structure of hNHE1.