A series of 80 compounds, primary, secondary, and tertiary amines and quaternary ammonium and bisammonium salts, most of them synthesized as potential choline or ethanolamine analogs, were tested against the in vitro growth of Plasmodium falciparum, the human malaria parasite. They were active over the 10(-3)-10(-8) M concentration range. A structure-activity relationship study was carried out using autocorrelation vectors as structural descriptors, and multidimensional analysis. Principal component analysis, ascending hierarchical classification, and stepwise discriminant analysis showed that both the size and shape of the molecule were essential for antimalarial potency, making the lipophilicity and electronegativity distribution in the molecular space essential. Using the autocorrelogram describing the molecular shape and the electronegativity distribution on the molecular graph, 98% of the molecules were correctly classified either as poorly active or active with only three explanatory variables. The most active compounds were quaternary ammoniums salts whose nitrogen atom had only one long lipophilic chain of 11 or 12 methylene groups (E5, E6, E10, E13, E20, E21, E22, E23, F4, F8), or the bisammoniums whose polar heads were linked by linear alkyl chains of 10 to 12 carbon atoms (G4, G23). The hydroxyethyl group of choline was not very beneficial, whereas the charge and substitutions of nitrogen (aimed at increasing lipophilicity) were essential for optimal interactions. A crude topographic model of the ligand (choline) binding site was thus drawn up.

Seventy-seven potential analogs of phospholipid polar heads, choline and ethanolamine, were evaluated in vitro as inhibitors ofPlasmodium falciparum growth. Their IC50 ranged from 10−3 to 10−7 mol/L. Ten compounds showed similar antimalarial activity when tested against three different parasite strains (2 chloroquine-sensitive strains and 1 chloroquine-resistant strain). Compounds showing marked antimalarial activity were assayed for their effects on phospholipid metabolism. The most active compounds (IC50 of 1 to 0.03 μmol/L) were inhibitors of de novo phosphatidylcholine (PC) biosynthesis from choline. For a series of 50 compounds, there was a close correlation between impairment of phospholipid biosynthesis and inhibition of in vitro malaria parasite growth. High choline concentrations caused a marked specific shift in the curves for PC biosynthesis inhibition. Concentrations inhibiting 50% PC metabolism from choline were in close agreement with the Ki of these compounds for the choline transporter inPlasmodium knowlesi-infected erythrocytes. By contrast, measurement of the effects of 12 of these compounds on rapidly dividing lymphoblastoid cells showed a total absence of correlation between parasite growth inhibition and human lymphoblastoid cell growth inhibition. Specific antimalarial effects of choline or ethanolamine analogs are thus likely mediated by their alteration of phospholipid metabolism. This indicates that de novo PC biosynthesis from choline is a very realistic target for new malaria chemotherapy, even against pharmacoresistant strains.