How selection was made for a new antimalarial drug candidate : from AM1 to MG3, two novel series of 4-aminoquinoline derivatives

Introduction Within AntiMal, our studies were focused on novel 4-aminoquinolines bearing a terminal bulky bicyclic basic moiety able to overcome parasite resistance. Among a first series, the quinolizidinylmethyl derivative AM1 showed potent in vitro and in vivo antimalarial activity. AM1 is a semi-synthetic product derived from (-)-lupinine (an alkaloid extracted from Lupinus luteus) which is expensive and available in amount not sufficient for market purposes. Two alternatives strategies were then employed : 1. AM1 was totally synthesised as racemic AM1, and consequently, the synthesis and the study of the (+)-enantiomer of AM1 was performed; 2. the design and synthesis of analogs of AM1, not chiral and cheaper was done. By replacing the quinolizidine nucleus of AM1 with a pyrrolizidine moiety we obtained MG3, a compound easily to be synthesized, cheaper and achiral. Methods Racemic lupinine was synthesized and the enantiomers separated by lipase catalyzed kinetic resolution of the racemate, with good yield and high optical purity. Then (±) AM1, and (+) AM1 were obtained in a 4 steps synthesis (Sparatore A. et al. 2005). The pyrrolizidinyl derivatives (MG2 and MG3) were easily obtained through few synthetic steps (Sparatore A. et al 2008). All compounds were assayed in vitro and in vivo for antimalarial activity and for cytotoxicity against normal human cells. In addition, the binding/internalization of AM1 and MG3 to normal human red blood cells (RBC) was measured by spectrofluorimetry and their effect on RBC membrane stability and RBC lysis was studied using the two compounds alone or in the presence of 10 µM heme. CQ was used as control. Results By comparing the synthetic process, the synthesis of MG3 resulted cheaper, easier and more suitable for industrial scale up than that of AM1. Both compounds were highly effective in vitro against multidrug resistant strains of P. falciparum (IC50 10-15 nM), not toxic and able to significantly reduce the parasitaemia in the mouse P. berghei model. MG3 binding to RBC was dose-dependent, linearly related to the amount of compound and similar to that of CQ. MG3 was not haemolytic per se up to the concentration of 100 µM and, differently from CQ or AM1 (Omodeo Salè et al. 2009) did not enhance the haemolytic activity of heme. Discussion The in vitro/ in vivo results, the good therapeutic index and the chemical feasibility allowed us to select MG3 as a suitable antimalarial drug candidate worthy of further development. The ANTIMAL-EU18834 support is acknowledged