Search

To ascertain the structure–activity relationship of the core 1,2,4-trioxolane substructure of dispiro ozonides OZ277 and OZ439, we compared the antimalarial activities and ADME profiles of the 1,2-dioxolane, 1,2,4-trioxane, and 1,2,4,5-tetraoxane isosteres. Consistent with previous data, both dioxolanes had very weak antimalarial properties. For the OZ277 series, the trioxane isostere had the best ADME profile, but its overall antimalarial efficacy was not superior to that of the trioxolane or tetraoxane isosteres. For the OZ439 series, there was a good correlation between the antimalarial efficacy and ADME profiles in the rank order trioxolane > trioxane > tetraoxane. As we have previously observed for OZ439 versus OZ277, the OZ439 series peroxides had superior exposure and efficacy in mice compared to the corresponding OZ277 series peroxides.

Thirty-three N-acyl 1,2,4-dispiro trioxolanes (secondary ozonides) were synthesized. For these ozonides, weak base functional groups were not required for high antimalarial potency against Plasmodium falciparum in vitro, but were necessary for high antimalarial efficacy in Plasmodium berghei-infected mice. A wide range of LogP/D(pH)(7.4) values were tolerated, although more lipophilic ozonides tended to be less metabolically stable.

The structure and stereochemistry of the cyclohexane substituents of analogues of arterolane (OZ277) had little effect on potency against Plasmodium falciparum in vitro. Weak base functional groups were not required for high antimalarial potency, but they were essential for high antimalarial efficacy in P. berghei-infected mice. Five new ozonides with antimalarial efficacy and ADME profiles superior or equal to that of arterolane were identified.

In four or five chemical steps from the 1,2,4-trioxane artemisinin, a new series of 23 trioxane dimers has been prepared. Eleven of these new trioxane dimers cure malaria-infected mice via oral dosing at 3 x 30 mg/kg. The clinically used trioxane drug sodium artesunate prolonged mouse average survival to 7.2 days with this oral dose regimen. In comparison, animals receiving no drug die typically on day 6-7 postinfection. At only 3 x 10 mg/kg oral dosing, seven dimers prolong the lifetime of malaria-infected mice to days 14-17, more than double the chemotherapeutic effect of sodium artesunate. Ten new trioxane dimers at only a single oral dose of 30 mg/kg prolong mouse average survival to days 8.7-13.7, and this effect is comparable to that of the fully synthetic trioxolane drug development candidate OZ277, which is in phase II clinical trials.

RBx11160 (OZ277) is a promising antimalarial drug candidate that Ranbaxy Laboratories Limited and Medicines for Malaria Venture (MMV) are currently developing as a fixed combination with piperaquine. Here, we describe the in vitro (Plasmodium falciparum) and in vivo (Plasmodium berghei) activities of piperaquine in combination with RBx11160 and artemether. In vitro, both combinations demonstrated a slight tendency towards antagonism with mean sums of fractional inhibitory concentrations (mean Sigma FICs) of 1.5. In vivo, piperaquine and artemether were borderline antagonistic (mean Sigma FIC of 1.4). However, an additive in vivo interaction of piperaquine and RBx11160 (mean Sigma FIC of 1.1) was identified, suggesting that a RBx11160-piperaquine combination therapy in humans should allow each molecule to exert its full antimalarial effect.

With an aim to identify a dispiro-1,2,4-trioxolane with high oral activity and good physicochemical properties, 27 derivatives of an achiral piperidine trioxolane were synthesized; most were potent antimalarial peroxides with IC(50)s ranging from 0.20 to 7.0 ng/mL. The oral efficacies of two of these were superior to artesunate and comparable to artemether. The attractive chemical simplicity of these compounds is balanced only by an apparent metabolic susceptibility.

Based on the structures of several lipophilic trioxolane antimalarial prototypes, we set out to determine which functional groups were associated with good antimalarial profiles and identify more polar (lower LogP/LogD) lead compounds with good physicochemical properties. More lipophilic trioxolanes tended to have better oral activities than their more polar counterparts. Trioxolanes with a wide range of neutral and basic, but not acidic, functional groups had good antimalarial profiles.

Fourteen spiro- and dispiro-1,2-dioxolanes were synthesized by peroxycarbenium ion annulations with alkenes in yields ranging from 30% to 94%. Peroxycarbenium ion precursors included triethylsilyldiperoxyketals and -acetals derived from geminal dihydroperoxides and from a new method employing triethylsilylperoxyketals and -acetals derived from ozonolysis of alkenes. The 1,2-dioxolanes were either inactive or orders of magnitude less potent than the corresponding 1,2,4-trioxolanes or artemisinin against P. falciparum in vitro and P. berghei in vivo. In reactions with iron(II), the predominant reaction course for 1,2-dioxolane 3a was two-electron reduction. In contrast, the corresponding 1,2,4-trioxolane 1 and the 1,2,4-trioxane artemisinin undergo primarily one-electron iron(II)-mediated reductions. The key structural element in the latter peroxides appears to be an oxygen atom attached to one or both of the peroxide-bearing carbon atoms that permits rapid beta-scission reactions (or H shifts) to form primary or secondary carbon-centered radicals rather than further reduction of the initially formed Fe(III) complexed oxy radicals.

This paper describes the discovery of synthetic 1,2,4-trioxolane antimalarials and how we established a workable structure-activity relationship in the context of physicochemical, biopharmaceutical, and toxicological profiling. An achiral dispiro-1,2,4-trioxolane (3) in which the trioxolane is flanked by a spiroadamantane and spirocyclohexane was rapidly identified as a lead compound. Nonperoxidic 1,3-dioxolane isosteres of 3 were inactive as were trioxolanes without the spiroadamantane. The trioxolanes were substantially less effective in a standard oral suspension formulation compared to a solubilizing formulation and were more active when administered subcutaneously than orally, both of which suggest substantial biopharmaceutical liabilities. Nonetheless, despite their limited oral bioavailability, the more lipophilic trioxolanes generally had better oral activity than their more polar counterparts. In pharmacokinetic experiments, four trioxolanes had high plasma clearance values, suggesting a potential metabolic instability. The toxicological profiles of two trioxolanes were comparable to that of artesunate.

The discovery of artemisinin more than 30 years ago provided a completely new antimalarial structural prototype; that is, a molecule with a pharmacophoric peroxide bond in a unique 1,2,4-trioxane heterocycle. Available evidence suggests that artemisinin and related peroxidic antimalarial drugs exert their parasiticidal activity subsequent to reductive activation by haem, released as a result of haemoglobin digestion by the malaria-causing parasite. This irreversible redox reaction produces carbon-centred free radicals, leading to alkylation of haem and proteins (enzymes), one of which--the sarcoplasmic-endoplasmic reticulum ATPase PfATP6 (ref. 7)--may be critical to parasite survival. Notably, there is no evidence of drug resistance to any member of the artemisinin family of drugs. The chemotherapy of malaria has benefited greatly from the semi-synthetic artemisinins artemether and artesunate as they rapidly reduce parasite burden, have good therapeutic indices and provide for successful treatment outcomes. However, as a drug class, the artemisinins suffer from chemical (semi-synthetic availability, purity and cost), biopharmaceutical (poor bioavailability and limiting pharmacokinetics) and treatment (non-compliance with long treatment regimens and recrudescence) issues that limit their therapeutic potential. Here we describe how a synthetic peroxide antimalarial drug development candidate was identified in a collaborative drug discovery project.