In vitro activity of anti-malarial ozonides against an artemisinin-resistant isolate

Malaria is one of the most important tropical diseases resulting in 214 million new cases and an estimated 438,000 malaria deaths worldwide in 2015 [1]. The discovery of artemisinin in the 1970s was an important step forward in anti-malarial drug therapy and was recognized with the Nobel Prize in Physiology or Medicine in 2015 [2, 3]. Artemisinin and its semi-synthetic derivatives, such as dihydroartemisinin (DHA) (Fig. 1), artesunate and artemether, contain a unique sesquiterpene lactone peroxide (1,2,4-trioxane) structure and artemisinin-based combination therapy (ACT) represents the current first-line treatment of uncomplicated Plasmodium falciparum malaria [4‐6]. Since the starting material artemisinin is a natural product, its production is limited to the availability of the plant [4, 7], although several total syntheses of artemisinin have been described [8]. In 2004, Vennerstrom et al. reported the discovery of a completely synthetic peroxide anti-malarial containing a 1,2,4-trioxolane (ozonide) pharmacophore named OZ277 (arterolane) (Fig. 1) with anti-malarial activity comparable to the artemisinin derivatives [9, 10]. In combination with piperaquine, arterolane was registered for anti-malarial combination therapy in India in 2011 [11‐14]. The next-generation ozonide, OZ439 (artefenomel) (Fig. 1), exhibits an increased pharmacokinetic half-life and good safety profile and is now being tested in phase IIb clinical trials [12, 14‐17].

The iron-dependent alkylation hypothesis is one of the proposed modes of action of artemisinin and synthetic peroxides [18‐21] where the peroxide is thought to be activated by the reductive cleavage in the presence of ferrous haem (or free Fe(II) derived from haem) released as a by-product of haemoglobin digestion in the food vacuole [20, 22‐27]. Thereby carbon-centred radicals are generated, which then alkylate haem and parasite proteins [28‐33]. The interaction of the artemisinin derivatives or ozonides with parasite targets is irreversible [31, 34]. Although the semi-synthetic artemisinins are highly effective, prolonged parasite clearance times were first reported along the Thai–Cambodian border in 2006, suggesting an emerging artemisinin resistance phenotype [35]. Today, delayed parasite clearance following treatment with artemisinin derivatives has been observed across Southeast Asia [36‐41]. It was found that mutations in the Kelch 13 propeller domain are associated with ring-stage parasites entering a quiescent state with delayed parasite clearance after exposure to artemisinins [41‐45]. When 50% inhibitory concentrations (IC₅₀) were measured using conventional methods such as the [³H] hypoxanthine incorporation assay [46], no difference was observed between artemisinin-resistant and -susceptible strains after treatment with artemisinin or its derivatives [47‐50]. In an effort to correlate the delayed parasite clearance observed in vivo with in vitro parasite survival, Witkowski et al. [48, 49] developed a ring-stage survival assay (RSA_0–3h) that exploited the differences in susceptibility observed between wild-type and K13 mutants at the early ring stage of the asexual blood cycle following a short pulse of artemisinin treatment. In the RSA_0–3h, synchronized young ring stage parasites (0–3 h old) are exposed to drugs for 6 h, and then cultured in drug free culture medium for 66 h before relative growth is determined by microscopic analysis [48, 49]. Since the structural analogies between artemisinins and ozonides (Fig. 1) suggest that they share similar modes of action, and thus some level of cross resistance [9, 10, 51, 52], the per cent survival of an artemisinin-resistant clinical isolate (Cam3.I^R539T) treated with DHA, OZ277, OZ439, and four additional next-generation ozonides (Fig. 1) using the RSA_0–3h as described by Witkowski et al. [48, 49] was evaluated. Additionally, a sub-set of these compounds was tested in the RSA_0–3h described by Xie et al. [53] that also uses tightly synchronized ring-stage cultures, but allows the assay to be performed routinely within a convenient time-frame.

The per cent survival of parasites exposed to a concentration range of DHA and six different ozonides (Fig. 1) was determined using the artemisinin-resistant P. falciparum Cambodian isolate Cam3.I^R539T. As expected, DHA exposure gave a high survival rate ranging from 74 to 33% at concentrations of 49 and 700 nM, respectively (Fig. 2), which is comparable to the observed survival value of 40% at 700 nM published previously [44]. In contrast, when tested at 700 nM, the two ozonides OZ277 and OZ439 showed an approximate 18- to 45-fold increase in potency compared with DHA (Fig. 2). Full and equal potency was observed when DHA, OZ277 and OZ439 were tested in parallel in the RSA_0–3h using the artemisinin-sensitive strain NF54 (Additional file 1: Table S1). At the lowest concentration (49 nM), OZ277 had poor activity, showing a similar per cent survival to that of DHA, whereas OZ439 was still about fivefold more potent. A possible explanation for OZ439 being more potent than OZ277 could be related to its improved stability in blood as previously described [15]. In those studies, OZ277 or OZ439 were incubated at 37 °C in P. falciparum-infected human blood. After 2 h more than 90% of OZ277 was degraded, whereas OZ439 was found to be about 10–20× more stable. A similar and more recent study found similar differences in stability for OZ277 and OZ439 [56]. The same compounds were also tested in a more convenient variation of the standard RSA_0–3h that uses synchronized ring-stage cultures that can be easily produced during normal working hours [53]. As shown in Table 1, this alternative synchronization method gave results that were comparable to those obtained using the standard RSA_0–3h.

To investigate further the level of cross-resistance between DHA and the ozonides, four additional next-generation ozonides (OZ493, OZ609, OZ655, OZ657) (Fig. 1) were tested against the Cam3.I^R539T parasites. While all six ozonides had a similar IC₅₀ value using a conventional 72-h [³H] hypoxanthine incorporation assay (Additional file 1: Table S2), the RSA_0–3h showed that OZ493, OZ609 and OZ655 were highly potent and completely inhibited the growth of the artemisinin-resistant isolate at the two highest concentrations tested (Fig. 2). At the lowest concentration, potency was comparable to that for OZ439. The overall potency of OZ657 was comparable to that of OZ277.

The RSA_0–3h was recently developed to provide an in vitro correlate of the longer in vivo parasite clearance times observed after artemisinin treatment in Southeast Asia, which is widely interpreted as a sign of potential artemisinin resistance [57, 58]. Provided that the RSA_0–3h does indeed predict the potency of compounds against artemisinin-resistant parasites in malaria patients, the here described data suggest that all of the tested ozonides are highly potent against isolates such as P. falciparum Cam3.I^R539T. These data are in line with the recent clinical observation that the parasite clearance rate following OZ439 treatment is not significantly affected by resistance-associated mutations in the Kelch 13 propeller region [17] and the recent data published by Siriwardana et al. [59], which showed no reduced susceptibility of OZ439 in a different delayed clearance phenotype parasite (Cam3.II) in vitro.

In the traditional RSA_0–3h, as well as a more convenient variation of the original method, all of the tested ozonides, were highly potent against the artemisinin-resistant isolate P. falciparum Cam3.I^R539T in contrast to results for DHA. These data indicate that artemisinin-resistant P. falciparum infections could be successfully treated with ozonide anti-malarial drugs.