MED DOI: 10.1002/cmdc.200700108 The Fe2+-Mediated Decomposition, PfATP6 Binding, and Antimalarial Activities of Artemisone and Other Artemisinins: The Unlikelihood of C-Centered Radicals as Bioactive Intermediates Richard K. Haynes,*[a] Wing Chi Chan,[a] Chung-Man Lung,[a] Anne- Catrin Uhlemann,[b] Ursula Eckstein,[b] Donatella Taramelli,[c] Silvia Parapini,[c] Diego Monti,[d] and Sanjeev Krishna*[b] The results of Fe2+ -induced decomposition of the clinically used facile extrusion of Fe2+ and collapse to benign isomerization artemisinins, artemisone, other aminoartemisinins, 10-deoxoarte- products. The propensity towards the formation of radical misinin, and the 4-fluorophenyl derivative have been compared marker products and intermolecular radical trapping have no re- with their antimalarial activities and their ability to inhibit the lationship with the in vitro antimalarial activities of the artemisi- parasite SERCA PfATP6. The clinical artemisinins and artemisone nins and trioxolane. Desferrioxamine (DFO) attenuates inhibition decompose under aqueous conditions to give mixtures of C radi- of PfATP6 by, and antagonizes antimalarial activity of, the aque- cal marker products, carbonyl compounds, and reduction prod- ous Fe2+-susceptible artemisinins, but has no overt effect on the ucts. The 4-fluorophenyl derivative and aminoartemisinins tend aqueous Fe2+-inert artemisinins. It is concluded that the C radi- to be inert to aqueous iron(II) sulfate and anhydrous iron(II) ace- cals cannot be responsible for antimalarial activity and that the tate. Anhydrous iron(II) bromide enhances formation of the car- Fe2+-susceptible artemisinins may be competitively decomposed bonyl compounds and provides a deoxyglycal from DHA and en- in aqueous extra- and intracellular compartments by labile Fe2+ , amines from the aminoartemisinins. Ascorbic acid (AA) acceler- resulting in some attenuation of their antimalarial activities. In- ates the aqueous Fe2+-mediated decompositions, but does not terpretations of the roles of DFO and AA in modulating antima- alter product distribution. 4-Oxo-TEMPO intercepts C radicals larial activities of the artemisinins, and a comparison with anti- from a mixture of an antimalaria-active trioxolane, 10-deoxoarte- malarial properties of simple hydroperoxides and their behavior misinin, and anhydrous iron(II) acetate to give trapped products towards thapsigargin-sensitive SERCA ATPases are presented. The in 73% yield from the trioxolane, and 3% from the artemisinin. general basis for the exceptional antimalarial activities of artemi- Artemisone provides a trapped product in 10% yield. Thus, in line sinins in relation to the intrinsic activity of the peroxide within with its structural rigidity, only the trioxolane provides a C radical the uniquely stressed environment of the malaria parasite is eminently suited for intermolecular trapping. In contrast, the thereby adumbrated. structural flexibility of the C radicals from the artemisinins allows Introduction [a] Prof. Dr. R. K. Haynes, W. C. Chan, C.-M. Lung Department of Chemistry, Open Laboratory of Chemical Biology The isolation and discovery of the antimalarial activity of arte- Institute of Molecular Technology for Drug Discovery and Synthesis misinin 1 represents one of the great events in medicine in the The Hong Kong University of Science and Technology latter half of the 20th Century.[1] Artemisinin and its derivatives Clear Water Bay, Kowloon, Hong Kong (PR China) dihydroartemisinin (DHA, 2), artesunate 3, and artemether 4[2–4] Fax: (+ 852)2358-1594 E-mail: [email protected] are now widely used for treatment of malaria caused by the [5] [b] Dr. A.-C. Uhlemann, Dr. U. Eckstein, Prof. Dr. S. Krishna most virulent parasite, Plasmodium falciparum. The pivotal Centre for Infection, Division of Cellular and Molecular Medicine problem in their use, that of a short pharmacological half-life, St. George’s Hospital, University of London SW17 0RE(UK) was uncovered by the Chinese and countered early on by ad- E-mail: [email protected] ministering artemisinin in combination with the pharmacologi- [c] Prof. Dr. D. Taramelli, S. Parapini cally robust mefloquine.[6] Consequently, artemisinins are now Dipartimento di Salute Pubblica-Microbiologia-Virologia Università di Milano, Via Pascal 36, 20133 Milan (Italy) routinely used in combination with longer half-life drugs,[7] and [d] Prof. Dr. D. Monti emphasis is now placed on the development of fixed combina- Institute of Science and Molecular Technology (ISTM) tions. Via Venezian 21, 20133 Milan (Italy) 1480 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemMedChem 2007, 2, 1480 – 1497 Decomposition Mechanism of Artemisinins droxylation of the periphery of the molecule syn to the perox- ide. This leaves the peroxide intact, as we first pointed out in 2001.[21] According to model studies, treatment of artemisinin with ferrous heme generates an artemisinin C radical, which undergoes an exceedingly rapid alkylation of the heme (of the order of a bond vibration).[22] As ligand exchange is required to free the putative C radical from the heme iron,[23] it cannot mi- grate away from the heme, and exclusively provides a heme– artemisinin adduct. Whilst the latter possesses antimalarial ac- tivity, this tends to be in the micromolar range in in vitro assays.[15,24] It has been suggested that the exogenously ap- plied adduct may not readily permeate the vacuolar mem- brane into the food vacuole.[15] Thus, this relatively weak activi- ty may be ascribed to the redox properties of the ferric heme In contrast to the universal acceptance of the utility of arte- nucleus within the adduct enhancing oxidative stress within misinins as the most potent and rapidly acting of all antimalari- the cytosol of the parasite, as discussed below. Current data al drugs, a seemingly intractable controversy surrounds the tend to indicate that the site of action of artemisinins is not manner in which these drugs are presumed to act as parasitici- within the food vacuole,[25] and it is established that artemisi- dal agents. It has been proposed that activity arises as a result nins do not inhibit hemozoin formation.[26] However, assertions of reductive or Fenton-like cleavage of the peroxide by Fe2+ to that artemisinins or artemisinin–heme adducts exert their ef- generate alkoxyl radicals, which either abstract hydrogen fects by inhibiting hemozoin formation,[24] thereby allowing atoms from the artemisinin periphery or undergo b-scission to buildup of ’toxic’ heme monomer in the food vacuole, or that provide carbon-centered radicals (C radicals) (Figure 1). The heme is both the ’activating agent’ and ’the target’ of artemisi- C radicals are held to alkylate proteins vital to parasite func- nin continue to appear.[15,16] The detection of radiolabeled arte- tion.[8–13] Formation of C radical intermediates in laboratory ex- misinin within artemisinin–heme adducts in minor amounts in periments is demonstrated by interception with thiols[10] or urine or spleen extracts of infected mice administered with an spin-trapping agents.[9,14] C Radicals are purportedly involved in excess of artemisinin,[27] whilst indicating unambiguous al- the alkylation of the heme when ferrous heme is used to though minor sequestration of artemisinins by heme, are indi- induce cleavage of the peroxide in artemisinin and ana- cative of routes of excretion of metabolites, and cannot be logues.[15,16] In the absence of trapping, the radical intermedi- taken at all as evidence of mechanism. Indeed, it is recognized ates proceed to form stable end products, such as 14 and 15 that competitive sequestration of artemisinins by heme, as from artemisinin (Figure 1 and below),[9,17, 18] which thereby takes place in the foregoing experiment, attenuates their anti- serve as markers for the intercession of such radicals. The malarial activity.[28] products are at the same oxidation level as the starting artemi- The idea that iron is required for activation of artemisinins sinin; the Fe2+ cation acts as a catalyst, as it is regenerated appears to rely in large measure on reports that iron chelators during their formation.[9] However, Fe2+ is considered to such as desferrioxamine B (DFO) antagonize the antimalarial engage in a second-order reaction with artemisinin under activities of artemisinins in vitro.[29,30] DFO is a hydrophilic che- aqueous conditions.[19] The results nevertheless are better inter- lating agent with exceptional affinity for Fe3+. The Fe3+ chelate preted in terms of a catalytic reaction, first-order in artemisinin, ferrioxamine B (FO) possesses a stability constant of 1031, being accelerated by the excess of the catalyst. which is 1024 times greater than that of the Fe2+ chelate.[31] No- Whilst an artemisinin–heme interaction in one form or an- tably, DFO is parasiticidal at a concentration of ~15 mm against other within the food vacuole of the malaria parasite has been P. falciparum, a level which is about 60-fold lower than that re- held responsible for antimalarial activity,[15,16] the ’heme theory’ quired for a cytostatic effect against mammalian cells. DFO af- itself is contentious. Artemisinins are vigorous inducers of CYP fects late trophozoites and early schizonts, whereas artemisi- enzymes[20] in which the iron in the heme prosthetic group nins are effective parasiticidal agents against the early ring must cycle through the ferrous state in order to induce hy- stage of the parasite. The antagonism cannot be reproduced Figure 1. Fe2+-catalysed conversion