Comparative Antimalarial Activities of Six Pairs of 1,2,4,5-Tetraoxanes (Peroxide Dimers) and 1,2,4,5,7,8-Hexaoxonanes (Peroxide Trimers)

Abstract

Six tetraoxanes had 50% inhibitory concentrations in the range of 10 to 100 ng/ml against Plasmodium falciparum, whereas the corresponding hexaoxonanes had minimal antimalarial activity. The lack of iron-mediated reactivity of the hexaoxonanes may explain their low activity compared to the tetraoxanes, the latter of which are able to undergo iron(II)-mediated activation.

The antimalarial artemisinin (qinghaosu) contains a pharmacophoric peroxide bond within its 1,2,4-trioxane heterocycle. The complex structure of artemisinin is an incentive to identify more synthetically accessible antimalarial peroxides (10, 13, 15, 18). One of the most structurally simple classes of antimalarial synthetic peroxides are 1,2,4,5-tetraoxanes (2, 5, 6, 12, 16), as exemplified by WR 148999 (Fig. 1). Tetraoxanes such as WR 148999 (20) differ considerably in structure from artemisinin, are readily prepared in one step from substituted cyclohexanones, and possess good antimalarial activity, although they are significantly less active than the semisynthetic artemisinins.

FIG. 1.

Structures of artemisinin and WR 148999.

In this study, we disclose the comparative antimalarial activities of six pairs of 1,2,4,5-tetraoxanes (peroxide dimers) 1a-1f and 1,2,4,5,7,8-hexaoxonanes (peroxide trimers) 2a-2f (Fig. 2) to further elucidate the structure-activity-relationship of 1,2,4,5-tetraoxanes and, to our knowledge, record for the first time the antimalarial activity of 1,2,4,5,7,8-hexaoxonanes. The latter group of synthetic peroxides has been investigated primarily as synthetic precursors to macrocyclic lactones of perfumery interest (17). As previously described (7, 8), tetraoxanes 1a-1f and hexaoxonanes 2a-2f were prepared either by peroxidation of the corresponding ketone derivatives in H₂SO₄/CH₃CN or by ozonolysis of the corresponding ketone O-methyl oximes. 1,3-Dioxolane 3, the nonperoxidic isostere of tetraoxane 1d, was prepared by ketalization of cyclohexanone with 1,1-dimethanolcyclohexane (TsOH, refluxing toluene) in 35% yield after Kugelrohr distillation (melting point 53 to 56°C) and characterized by ¹H and ¹³C NMR and elemental analysis.

FIG. 2.

Structures of 1,2,4,5-tetraoxanes 1a to 1f, 1,2,4,5,7,8-hexaoxonanes 2a to 2f, and nonperoxidic isostere 3.

As previously described (6), in vitro and in vivo antimalarial activities were measured using the chloroquine-resistant K1 and chloroquine-sensitive NF54 strains of Plasmodium falciparum- and P. berghei-infected mice, respectively. Groups of three P. berghei-infected MORO mice were treated 1 day postinfection with 100-mg/kg oral and subcutaneous doses of compounds dissolved or suspended in a solubilizing 3% ethanol and 7% Tween 80 vehicle. Antimalarial activity was measured by determining the percent reduction in parasitemia on day 3 postinfection compared to an untreated control group.

The data in Table 1 show a clear demarcation in the antimalarial activities of tetraoxanes 1 and hexaoxonanes 2 against P. falciparum. Tetraoxanes 1 had 50% inhibitory concentrations (IC₅₀s) in the range of 10 to 100 ng/ml, whereas all hexaoxonanes had minimal antimalarial activity; only three of these (2a, 2d, and 2f) had IC₅₀s of <10,000 ng/ml. The previously reported IC_50s against the P. falciparum W2 and D6 clones for 1d (20 and 26 ng/ml) and 1e (68 and 100 ng/ml) (11) and against the P. falciparum FCB1 strain for 1a (48 ng/ml) and 1f (21 ng/ml) (21) are quite similar to those that we measured against the P. falciparum K1 and NF54 isolates. Of tetraoxanes 1, only 1f was as potent as WR 148999. However, each of tetraoxanes 1a to 1f was much less active than WR 148999 against P. berghei. Apparently, the two methyl groups flanking the central 1,2,4,5-tetraoxane heterocycle in WR 148999, which provide substantial steric hindrance to the peroxide bonds, are critical for good in vivo activity. Ketal 3, the nonperoxidic isostere of tetraoxane 1d, was devoid of antimalarial activity demonstrating that at least one of the two peroxide bonds of 1d (and other tetraoxanes) is required for antimalarial activity.

TABLE 1.

Activity of 1,2,4,5-tetraoxanes 1, 1,2,4,5,7,8-hexaoxonanes 2 and nonperoxidic isostere 3 against P. falciparum in vitro and P. berghei in vivo

Compounda	IC50 (ng/ml)b K1/NF54	Activity (%)c (p.o./s.c.)
1a	40/23	21/0
1b	71/88	47/5
1c	51/23	35/84
1d	30/44	0/26
1e	78/44	22/0
1f	14/11	60/56
2a	7400/5500	ND
2b	>10,000/>10,000	ND
2c	>10,000/>10,000	ND
2d	770/2700	ND
2e	>10,000/>10,000	ND
2f	2000/3200	ND
3	>10,000/>10,000	ND
WR 148999*	7.2/10	99.7/>99.99
Artemisinin*	2.8/3.4	98/>99.99

^a*, Data from Dong et al. (6).

^bMean from n = 2 to 3 determinations. Individual measurements differed by <50%.

^cMean from n = 3 determinations. Individual measurements differed by <10%. ND, not determined. Data are presented as values for peroral (p.o.)/subcutaneous (s.c.) administrations.

A long-standing hypothesis (9, 14, 15) to account for the antimalarial specificity of artemisinin is that the peroxide bond undergoes reductive activation by heme released by parasite hemoglobin digestion. This irreversible redox reaction produces carbon-centered free radicals or carbocations that alkylate heme or parasite proteins (9). Since the antimalarial activity of tetraoxanes 1 is peroxide bond dependent and hexaoxonanes 2 are without significant antimalarial activity, we wondered whether this observation could be explained in part by the relative susceptibilities of 1 and 2 to reaction with iron(II). To this end, we investigated the reaction of the tetraoxane/hexaoxonane pair 1d/2d with inorganic ferrous iron using previously established standardized conditions (3). In this experiment, pseudo-first-order reaction rate constants for 1d and 2d (0.03 mM) with FeSO₄ (3 mM) in 50% acetonitrile-water at 37°C (n = 3) under an argon atmosphere were determined. The experiment was corrected for nonspecific degradation in iron-free controls. Tetraoxane 1d underwent complete degradation in the presence of FeSO₄ with a pseudo-first order rate constant (k) of 0.38 ± 0.04 h⁻¹. In contrast, hexaoxonane 2d did not undergo any significant degradation in the presence of FeSO₄ over the 15 h time course studied (k < 0.007 h⁻¹ − the upper limit of the 95% confidence interval). The lack of iron-mediated reactivity of 2d may explain the low antimalarial activity of the hexaoxonanes 2 compared to the tetraoxanes 1, the latter of which are able to undergo iron(II)-mediated activation. For comparison, the pseudo-first-order rate constant (k) for artemisinin is 0.054 ± 0.006 h⁻¹ (3).

The difference in iron-mediated reactivity between 1d and 2d can be explained by considering their peroxide bond steric accessibility as shown by molecular modeling using Spartan'06 (Wavefunction, Inc.) (Fig. 3). The peroxide bonds in 1d appear to be fairly accessible, whereas the peroxide bonds in 2d are shielded by the three cyclohexane rings so that an iron(II) complex would be sterically prohibited from undergoing inner-sphere electron transfer with the peroxide bond antibonding σ* (LUMO) orbitals. Similar arguments have been put forth to explain the low antimalarial activity of sterically congested 1,2,4-trioxolanes (4), 1,2,4-trioxanes (19), and 1,2,4-trioxepanes (1).

FIG. 3.

Modeled structures of 1,2,4,5-tetraoxane 1d and 1,2,4,5,7,8-hexaoxonane 2d depicted with LUMOs (blue) mapped on isodensity surfaces.