Design and synthesis of antimalarial endoperoxides
More than 600 million people worldwide are infected with malaria (WHO 2004), and an average of 1-2 million die every year, most below the age of five. Several non-ACTs (Artemisinin-based Combination Therapies) have been used to fight this infectious disease, but only ACTs have continued to show efficacy effective with no reported cases of resistance.
Artemisinin 1a (0.01-0.8% dry weight from leaves and flowers of Artemisia annua L., Phytochemistry, 1996, 43, 993-1010. ) and its semi-synthetic derivatives 1b-e (first generation) are available for oral, intramuscular, rectal and intravenous administration, while new, fully synthetic endoperoxides (second generation) are being developed and have been associated with good activity in both in vitro systems and animal models.
Endoperoxides 2a-c, with a 2,3-dioxabicyclo[3.3.1]nonane framework, were synthesized in eight steps starting from the commercially available (R)-limonene oxide 3 in 0.4% (2a), 0.6% (2b) and 0.8% (2c) overall yields. The key steps in this synthesis were a four-component sequential thiol-olefin co-oxygenation (TOCO) reaction to obtain 7, and a Pummerer-type oxidative desulphurization to yield aldehyde 11. The FeII-catalyzed fragmentation of compounds in this class can generate chalcones that are effective antimalarial cysteine protease inhibitors, as well as potentially cytotoxic carbocations that arise via FeIII-oxidation of the resulting secondary C-centred radicals (Angew. Chem. Int. Ed. 2004, 43, 4193-4197. ) .
Based on previous experimental results, the Posner group has synthesized potential antimalarial drug candidates such as 12a-d (two steps, in 41% (12a) and 40% (12b) overall yields) and 13 (three steps, in 20% (α) and 33% (β) overall yields).
These prodrugs can produce highly reactive intermediates (e.g. oxy and carbon radicals, high-valent iron-oxo species) and other long lived neutral electrophiles (e.g. epoxides, aldehydes, dicarbonyl compounds) when activated by ferrous iron.
Recently prepared oil- and water-soluble C-10 phenoxy analogues (1: R1 = H, R2 = O-aryl; one and two steps from 1b, respectively) were shown to be superior to 1e, the drug of choice for most cases of malaria in the developing world, and are under further evaluation (Acc. Chem. Res. 2004, 37, 397-404. ) .
Trioxolane 14 (OZ277 or RBx-11160), currently under clinical evaluation, presents a simple, economical (three steps in 60% overall yield) and scalable synthesis, and does not suffer from safety problems such as those that caused Hoffmann-LaRoche to discontinue development of the antimalarial drug candidate arteflene (RO-42-1611) (15) (Exp. Opin. Ther. Patents 1998, 11, 1487-1493. Link).
The key step was the preparation of 16 by a coozonolysis of O-methyl 2-adamantanone oxime with a 4-substituted cyclohexanone. Neurotoxicity in animal models has raised concerns about ACTs, but tests have established the absence of quantifiable concentrations in the brain tissue over an 18h sampling period, and no mutagenic potential or other minor clinic findings, gastric irritation, hepatocellular hypertrophy, renal tubular changes, and atrophy of lymphatic tissue have been observed at high doses (Nature 2004, 430, 900-904. ) .
Large-scale and low-cost syntheses of endoperoxides 17, 18 and 19 (selected examples, three steps in 13 % (17), 16% (18), and 22% (19) overall yields) have been developed. The key step was a [4+2] cycloaddition of singlet oxygen with substituted 1,3-butadienes, with a further epoxidation required to mimic the artemisinin (1a) 1,2,4-trioxane ring (J. Med. Chem. 2004, 47, 1833-1839. ) .
a) [3H]-hypoxanthine incorporation assay (Antimicrob. Agents Chemother.1979, 16, 710).
c) Chloroquine-sensitive (an airport strain of unknown origin)
d) Chloroquine-sensitive (Cell 1988, 57, 921-930. )