ABSTRACT
The in vivo antimalarial efficacies of two phosphatidylinositol 4-kinase (PI4K) inhibitors, a 3,5-diaryl-2-aminopyrazine sulfoxide and its corresponding sulfone metabolite, were evaluated in the NOD-scid IL2Rγnull (NSG) murine malaria disease model of Plasmodium falciparum infection. We hypothesized that the sulfoxide would serve as a more soluble prodrug for the sulfone, which would lead to improved drug exposure with oral dosing. Both compounds had similar efficacy (90% effective dose [ED90], 0.1 mg kg–1 of body weight) across a quadruple-dose regimen. Pharmacokinetic profiling revealed rapid sulfoxide clearance via conversion to sulfone, with sulfone identified as the major active metabolite. When the sulfoxide was dosed, the exposure of the sulfone achieved was as much as 2.9-fold higher than when the sulfone was directly dosed, thereby demonstrating that the sulfoxide served as an effective prodrug for the treatment of malaria.
INTRODUCTION
Malaria, a mosquito-borne disease caused by infection of Plasmodium parasites, is one of the leading causes of death in sub-Saharan Africa. The WHO estimates that there were approximately 216 million cases of malaria worldwide, with 445,000 deaths, in 2016 (1). Public health measures and combination chemotherapy with artemisinin derivatives and partner compounds have largely succeeded in reducing morbidity and mortality over the last decade, but clearly, more progress needs to be made. Worrying reports are emerging around artemisinin-resistant parasites, which has become particularly troublesome since artemisinin-based combination therapy (ACT) regimens are currently considered the optimal standard of care (2–4). An urgent need therefore exists for treatment modalities that will overcome resistance, such as the development of novel and more efficacious antimalarial drugs and new drug combinations.
We have previously reported orally active antimalarial agents based on the 2-aminopyridine scaffold (5, 6). This work led to the identification of MMV390048, which was potent across the plasmodial life cycle except against hypnozoites, and which is currently in phase 2 clinical development (7). MMV390048 operates by inhibition of Plasmodium phosphatidylinositol 4-kinase (PI4K), a novel mode of action relative to all other clinical antimalarial agents. As MMV390048 offers many advantages to current antimalarial therapies, including high activity against all resistant phenotypes (50% inhibitory concentration [IC50], 17 to 26 nM) studied so far, and given the inherent risk of failure during clinical development, the identification of back-up or newer-generation PI4K inhibitors thus becomes imperative (7). Further structure-activity relationship (SAR) exploration around MMV390048 led to the identification of closely related 2-aminopyrazines that maintain potent antimalarial activity and other favorable drug metabolism and pharmacokinetic (DMPK) properties, including half-lives greater than 4 h in rats (8). The morphological effects of the aminopyrazines on Plasmodium spp. are likely similar to those of MMV390048 (7), though this has not been investigated. The lead aminopyrazine, a sulfone (Fig. 1), illustrated promising in vitro antiplasmodial and remarkable in vivo antimalarial activity with upwards of 99% reduction of parasitemia on 4-day quadruple dosing at 1 mg kg–1 of body weight in a Plasmodium berghei mouse model of infection (8). However, its poor solubility was identified as a key risk associated with its progression as a preclinical development candidate (8). Hence, an analogue program commenced wherein highly active more-soluble analogues with favorable pharmacokinetic attributes were identified (9). As another approach, we recognized that the corresponding sulfoxide analogue was more soluble and had similar in vitro activity, with an IC50 of 15.9 nM against NF54 Plasmodium falciparum compared to the sulfone (IC50, 8.4 nM). Furthermore, we hypothesized that the sulfoxide would serve as a prodrug for the sulfone, which would lead to improved drug exposure on oral dosing due to the higher solubility, as has been described for other sulfoxides (10, 11). In this paper, we extend the work in this series by testing the aforementioned hypotheses in an in vivo humanized model of malaria.
Structures of MMV390048, the aminopyrazine sulfone, and its corresponding sulfoxide.
A promising approach to test efficacy against human malaria is the use of a murine disease model utilizing immunocompromised NSG mice developed at GlaxoSmithKline (GSK) Tres Cantos (12, 13) and now implemented at the University of Cape Town’s Drug Discovery and Development Centre (H3D). NSG mice are able to be intraperitoneally engrafted with human red blood cells, resulting in a gradual increase in human erythrocytes in peripheral circulation, likely via the uptake by the lymphatic system. The mice are then intravenously infected with P. falciparum that can specifically infect the available human erythrocytes (14). Relative to P. berghei infection models, being able to model P. falciparum infections using human erythrocytes in vivo offers the clear advantage that efficacy would better characterize the human disease condition.
In this paper, we use the NSG mouse model to achieve two main objectives, as follows: first, to determine whether administration of a more-soluble prodrug sulfoxide leads to a higher exposure of the corresponding sulfone relative to dosing the sulfone itself, and second, to determine whether this increase in exposure leads to better antimalarial efficacy.
RESULTS
In vitro antiplasmodial activity, physicochemical properties, and in vitro microsomal metabolism.The activities of the sulfoxide were comparable between the adapted Pf3D70087/N9 and PfNF54 strains (Table 1), and they were only marginally lower than those of the sulfone.
In vitro antiplasmodial activity, physicochemical properties, and in vitro microsomal metabolism data of sulfoxide and sulfone
As expected, the sulfoxide was more soluble and less lipophilic than the sulfone. However, the passive permeabilities of compounds were comparable. The sulfoxide was moderately stable in human (HLM) and mouse liver microsomes (MLM), while the sulfone was considerably more stable.
Pharmacokinetics of sulfoxide and sulfone.The in vivo PK of the sulfoxide and sulfone was determined in healthy male C57BL/6 mice (Fig. 2) upon intravenous (i.v.) dosing at 1.5 mg kg−1. The sulfoxide showed a short half-life (t½) (<1 h) and was rapidly cleared with high conversion to the sulfone as the main metabolite. The exposure (area under the concentration-time curve [AUC]) of the sulfone in whole blood over the course of the experiments was more than 15-fold higher than that of the parent sulfoxide (see Table S3 in the supplemental material).
Mean log circulating levels of the sulfoxide (bold lines), its sulfone metabolite (dotted lines), and sulfone only following i.v. administration of either sulfoxide or sulfone.
Metabolite identification.The main metabolite generated from the sulfoxide both in vitro and in vivo was the corresponding sulfone, with the only other metabolite being a minor dioxidation metabolite (see the supplemental material). The dioxidation metabolite was also detected at low levels from the sulfone both in vitro and in vivo. Microsomal incubations confirmed that metabolism of the sulfoxide and sulfone was enzyme and NADPH dependent.
Therapeutic efficacy of sulfoxide and sulfone.(i) Quadruple-dose regimen of sulfoxide and sulfone. Fig. 3 shows the parasitemia observed for the sulfoxide and sulfone following a quadruple-dose regimen against P. falciparum. In both cases, four consecutive doses of 0.5 mg kg−1 and 1.5 mg kg−1 translated into the highest efficacy, with 0.1 mg kg−1 being less effective and 0.05 mg kg−1 being ineffective, rendering results similar to those of the untreated control group. Interestingly, the highest dose of 10 mg kg−1 did not translate into the highest efficacy, though the exposures (peak concentration of drug in whole blood [Cmax] and AUC) were clearly higher (Table S4).
Percent parasitemia observed in the NSG model of P. falciparum infection. (a) Daily quadrupole oral dosing of sulfoxide. (b) Daily quadruple oral dosing of sulfone. (c) Single oral dose of sulfoxide. The result for chloroquine (CQ) is included in each graph.
Blood samples collected on the first day of treatment were used to determine the PK parameters for both the sulfoxide and sulfone in the infected NSG mice. The sulfoxide was rapidly absorbed (time to Cmax [Tmax], 0.375 to 2 h) at all doses, with the sulfone being detected as the main metabolite. The sulfone had a much higher exposure (Table S4), suggesting that the efficacy observed for the sulfoxide was mainly due to this conversion. The dioxidation metabolite of the sulfoxide (mono-oxidation of the sulfone) is considered too low level to be consequential for the observed efficacy. The absorption of the sulfone was considerably slower (Tmax, 4 to 7 h), and its exposure (AUC) was higher following oral administration of the sulfoxide, demonstrating the benefits of the prodrug approach. As the dose escalated, the ratio of sulfone to sulfoxide decreased; for example, at 1.5 mg kg−1, the ratio was 30, while at 10 mg kg−1, it was 16.
(ii) Single-dose regimen of the sulfoxide. The PK profiles of both the sulfoxide and its metabolite showed that significant levels of the sulfone were still in circulation 24 h after dosing. This would suggest that the sulfone metabolite accumulates during a quadruple-dose regimen of the sulfoxide in NSG mice. A single-dose study was therefore performed to assess the efficacy and PK of the sulfoxide. The PK profile is summarized in Table S5. The in vivo therapeutic efficacy against P. falciparum following a single dose of the sulfoxide is illustrated in Fig. 3. The single dose of 1.5 mg kg−1 translated to the highest efficacy, with higher doses leading to a lower drop in parasitemia, despite increases in exposure as the dose escalated. Similar reductions in parasitemia were seen for the single dose of 15 mg kg−1 compared to the four consecutive doses of 10 mg kg−1 (Fig. 3).
Exposure in healthy versus infected mice.The oral PK parameters of the sulfoxide were determined in healthy NSG and infected NSG mice, as well as in healthy C57BL/6 mice, at 1.5 mg/kg (Table 2). The combined AUCs of the sulfoxide and sulfone were similar in healthy C57BL/6 and healthy NSG mice, suggesting that absorption of the sulfoxide was comparable in the two strains. However, the sulfoxide-to-sulfone conversion was faster in the NSG mice, as evidenced by the lower sulfoxide Cmax and the higher sulfone-to-sulfoxide ratio. Exposures of the sulfoxide and sulfone were marginally lower in infected NSG mice, suggesting an effect of the infection on the pharmacokinetics of the compounds.
Comparison of pharmacokinetic parameters of the sulfoxide and its sulfone metabolite following oral administration of 1.5 mg kg−1 sulfoxidea
Dose-response relationship in the NSG malaria mouse model.To evaluate the ED90 (on day 7 following infection) of the compounds, dose-response experiments were conducted in the NSG mouse model of malaria at a quadruple-dose regimen of 0.05, 0.1, 0.5, 1.5, and 10 mg kg−1 administered daily for four consecutive days. An additional single-oral-dose experiment was carried out with the sulfoxide, at 0.5, 1.5, 5, 15, and 50 mg kg−1. In both series of experiments, dosing was initiated on day 3 following infection, and the percent parasitemia was evaluated on days 3 to 7 daily as the primary efficacy readout. The relationship between parasitemia and dose was described by a sigmoid dose-response model and was used to determine the ED90 and AUCED90 for the compounds (Table 3). Sigmoidal plots of parasitemia relative to dose and AUC are illustrated in Fig.S7 and S8 in the supplemental material. The ED90 value of 2.80 mg kg−1 obtained for chloroquine compared well with those found in previous publications (14). Both parent compounds displayed a maximum reduction in parasitemia compared to the untreated control groups, with ED90 values of 0.12 mg kg−1 for both the sulfoxide and sulfone following the quadruple-dose regimen. The average estimated daily exposures (AUCED90) required to reduce parasitemia by 90% compared to the untreated control group were 0.39 and 0.20 µg h ml−1 day−1 for the sulfoxide and sulfone, respectively. The single-dose study resulted in higher ED90 and AUCED90 values of 0.51 mg kg−1 and 1.33 µg h ml−1 over the 96-h period, respectively (Table 3 and Fig. S8). Single dosing of the sulfoxide therefore was demonstrated to be highly efficacious against P. falciparum in the model, though it was less efficacious than multiday dosing, as would be expected. In order to determine a pharmacokinetic/pharmacodynamic (PK/PD) driver responsible for the observed efficacy, we determined the goodness-of-fit profiles in the form of R2 values derived from the sigmoid dose-response model used to describe the data. However, no concise conclusion could be made from the R2 values due to the steepness of the slope.
Efficacy parameters for both compounds with log fit against P. falciparum Pf3D70087/N9
DISCUSSION
The NSG murine malaria model allows in vivo evaluation of the antimalarial efficacy of promising compounds. The use of P. falciparum, a human pathogen, allows a determination of pharmacodynamic parameters that can be used to guide the prediction of such parameters in humans. We used this model to investigate the efficacy of a sulfoxide, its sulfone metabolite, and the contribution of this conversion to antimalarial activity. The ED90 values of 0.12 mg kg−1 in the model for both the sulfoxide and sulfone were lower than those for published compounds in this series, namely MMV390048 (ED90, 0.57 mg kg−1) and UCT943 (ED90, 0.25 mg kg−1) (7, 24). Clearly, the higher activity in the model is due to the combination of multiple factors, including in vitro efficacy, protein binding, partitioning into red blood cells, and pharmacokinetics.
Metabolism of sulfoxides typically produces sulfones as metabolites, though several enzymes, including cytochrome P450s, also catalyze the reductive reaction to sulfides (25–27). Sulfoxides are typically more soluble than the corresponding sulfones and can therefore be exploited as prodrugs for the delivery of sulfones (10). As part of our efforts to develop new antimalarial compounds, we sought to investigate whether this prodrug approach could be used to deliver higher exposures of the sulfone than might be achieved by dosing the sulfone itself and to investigate the contribution of the sulfoxide and sulfone to the in vivo antimalarial efficacy.
There were no significant in vitro differences between the IC50s of the sulfoxide against the strains of P. falciparum tested, including against the Pf3D70087/N9 strain adapted for use in the NSG infection model. Furthermore, the main metabolite detected both in vitro and in vivo was the sulfone metabolite. As both the sulfoxide and sulfone are similarly active against P. falciparum in vitro, activity in an in vivo model of infection is likely to reflect the exposure of both compounds combined. However, the conversion kinetics of the sulfoxide into the sulfone was sufficiently rapid (AUC ratios of sulfone to sulfoxide were greater than 30, as seen in the i.v. PK with the C57BL/6 mice), such that the demonstrated in vivo activity was largely due to the sulfone. This clearly translated in a similar manner to oral dosing in the NSG mouse, though there was a measure of saturation for the conversion of the sulfoxide to the sulfone, with the sulfone-to-sulfoxide ratios decreasing as the dose escalated. The sulfone was also the main metabolite in human liver microsomes, suggesting that metabolism should be qualitatively similar in mice and humans. Overall, oral administration of the sulfoxide at all the doses utilized resulted in higher exposures of the sulfone than with administration of the sulfone itself (Fig. 4).
Exposure (AUC0–24) of the sulfoxide, sulfone metabolite, and sulfone when either the sulfoxide or the sulfone is administered orally to infected NSG mice.
For example, close to the ED90 of 0.12 mg kg−1 (that is, the 0.1 mg kg−1 dose), the AUC of the sulfone due to sulfoxide dosing was 1.8-fold higher than that seen with dosing the sulfone itself. At the highest 10-mg kg−1 doses, the differential was 2.9-fold. These results show that dose escalation of the sulfoxide led to a superscaling effect, whereby the higher doses delivered disproportionately higher exposures of the sulfone, while higher doses of the sulfone led to flat if not lower relative exposures. Hence, the sulfoxide indeed served well as a prodrug to deliver higher drug exposures. The lower solubility of the sulfone made it difficult to achieve higher exposures. The superscaling seen with dose escalation of the sulfoxide indicates that at least one important clearance mechanism for the sulfone was saturated. Since the sulfoxide and sulfone are similarly active against P. falciparum in vitro, it is well understood for precision of PK/PD analyses that total concentrations of sulfoxide plus sulfone would be slightly better to use for studies with the sulfoxide. In vivo, the sulfone illustrated a long half-life of 16 h. This suggests that appropriate dosing would allow for a sustained exposure, which is important to a potential single-dose eradication of the parasite.
The sulfoxide showed potent in vivo activity, with an ED90 of 0.12 mg kg−1 in the 4-day dose NSG model of infection. It was therefore equipotent to the sulfone, which displayed a similar ED90. The PK profile of the sulfoxide showed that it was rapidly cleared, with no levels detectable past 10 h at all doses. The AUC of the sulfone metabolite was 2- to 3-fold higher than the AUC of the parent compound and remained detectable for more than 24 h after dosing. The reason why the higher exposure resulting from dosing the sulfoxide relative to the sulfone did not lead to a measurable improvement in ED90 remains unclear. This could be due to the imprecise determination of the ED90 resulting from the sharp inflection of the dose-response and exposure-response curves. Alternatively, a hint to understanding PK/PD comes from the observation that though the 0.05-mg kg−1 dose of the sulfoxide and the 0.1-mg kg−1 dose of the sulfone had about equal AUC values, the sulfoxide performed less well, with no drop in parasitemia, relative to the sulfone with a 90% drop in parasitemia (Fig. 3). The Cmax was lower for the sulfoxide (9.6 µg ml−1) than for the sulfone (14.2 µM), and the Tmax was later (Table S4). It is possible that Cmax influences efficacy more strongly than AUC, accounting for the higher parasitemia of the 0.05-mg kg−1 dose than that of the 0.1-mg kg−1 dose. Clearly, a more extended dose fractionation study would help elucidate the primary driver of efficacy (28, 29).
Interestingly, we also observed that the highest dose in the quadruple-dosing experiments for both the sulfoxide and sulfone increased rather than reduced parasitemia relative to the optimal 0.5- to 1.5-mg kg−1 doses. The sulfoxide was also effective as a single dose, with an ED90 of 0.51 mg kg−1, and similarly showed an (albeit less pronounced) optimal dose of 1.5 mg kg−1. There is considerable interest in identifying drug regimens that would deliver single-dose cures (30), hence the value of comparing single- versus multiple-day dosing. Varying the doses and dosing intervals will further help delineate PK/PD relationships that are important toward maximizing ultimate efficacy in the clinical setting. However, the cause is unclear for the seemingly aberrant dose and exposure responses, a PK/PD discrepancy that has not been seen before with other compound classes of antimalarials (31–33). The expectation might be that efficacy would plateau with achievement of a maximum effect and there might be a limit to the speed of parasite clearance. Experimental variability once this plateau is reached does not explain this decreasing parasitemia result, as similar results were seen for both the sulfone and sulfoxide, with quadruple dosing versus single dosing, and for other PI4K-inhibiting aminopyridines and aminopyrazines. We speculate that there is involvement of a host factor that is inhibited by these PI4K inhibitors in the malaria parasite. This class of compounds is selective for Plasmodium over human PI4K as well as most other human lipid and protein kinases that are believed to be associated with intracellular signaling and trafficking and host defense (20). Binding to human lipid kinase PIP4K2C (phosphatidylinositol-5-phosphate 4-kinase, type II, gamma) has been demonstrated, but its expression in red blood cells is unknown, as is the consequence of its inhibition (7).
Conclusion.Following a quadruple-dose regimen and a single-dose regimen, the aminopyrazine sulfoxide was demonstrated to be highly active against P. falciparum in the NSG mouse model of malaria infection. PK profiling showed that the sulfoxide was converted to the corresponding equally active aminopyrazine sulfone. The higher levels of the sulfone relative to the sulfoxide, when sulfoxide is dosed, and its longer half-life suggest that the sulfone metabolite is responsible for the efficacy observed with the sulfoxide. Furthermore, dosing with the sulfoxide generated higher exposures of the sulfone metabolite than dosing with the sulfone, which demonstrates the pharmacokinetic effectiveness of the prodrug approach. This effectiveness is predicted to translate to humans based on the qualitative similarity in microsomal metabolisms between mouse and human liver microsomes. However, the prodrug-mediated increase in exposure did not lead to an improvement in vivo efficacy, as there are as-yet-unexplained limiting factors associated with the pharmacodynamic response observed with the levels of parasitemia. Within this context, it is acknowledged that the NSG model does have limitations compared to the human condition, not the least of which is the absence of an immune system. Hence, more investigations are needed to understand the PK/PD relationships associated with the sulfoxide, sulfone, and other similarly active PI4K-inhibiting antimalarial agents.
MATERIALS AND METHODS
Chemistry.The aminopyrazine sulfoxide and sulfone were synthesized from commercially available 2-aminopyrazine, as previously described (8).
In vitro efficacy.Compounds were screened against multidrug-resistant (PfK1) and -sensitive (PfNF54) strains of P. falciparum in vitro using the modified [3H]hypoxanthine incorporation assay (15). Additionally, the compounds were evaluated against the adapted strain of P. falciparum 3D7 harvested from the NSG model of infection as a comparator; this was done using the Plasmodium lactate dehydrogenase (pLDH) assay (16). All assays were performed in triplicate in two independent experiments.
In vitro absorption, distribution, metabolism, and excretion assay.Kinetic solubility, lipophilicity, metabolic stability, and parallel artificial membrane permeability assay (PAMPA) of the compounds were measured using standard protocols, as previously described (17).
Ethics statement.All studies and procedures were conducted with prior approval of the animal ethics committee of the University of Cape Town (approval numbers 013/028 and 014/028) in accordance with the South African National Standard (SANS 10386:008) for the Care and Use of Animals for Scientific Purposes (18), and guidelines from the Department of Health (19).
In vivo pharmacokinetics in healthy C57BL/6 mice.Before designing the NSG experiments, it was necessary to establish the pharmacokinetic profiles of both compounds in a generic mouse strain. These experiments were performed in C57BL/6 mice, which have been routinely used for other compounds (20–22). The sulfoxide and sulfone were administered orally and intravenously at a dose of 1.5 mg kg−1 (n = 3). For the oral administration, the compounds were formulated in 0.5% hydroxypropylmethyl cellulose (HPMC) containing 0.2% Tween 80. The intravenous doses were administered in dimethylacetamide-polyethylene glycol 400-polypropylene glycol (DMA-PEG400-PPG) (10:30:60). Blood samples were collected via tail bleeding at predetermined time points and stored at −80°C until liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis.
In vivo efficacy and pharmacokinetics.(i) In vivo antimalarial efficacy studies. The therapeutic efficacy of both compounds was evaluated using the NSG mouse model for malaria previously described (13, 14). Male NSG mice engrafted with human erythrocytes (approximately 60%) were infected with 20 × 106 P. falciparum Pf3D70087/N9 cells, a strain developed at GSK for proliferation in engrafted mice. Infections were conducted via intravenous injection (day 0). Treatment commenced on day 3 and ended on day 7 following infection.
The efficacy of the sulfoxide and sulfone was evaluated following a quadruple-dose regimen, administered as a single oral dose for four consecutive days. Additionally, the sulfoxide was evaluated following a single-dose regimen.
For the quadruple-dose regimen, the compounds were dosed as single doses for four consecutive days at 0.05, 0.1, 0.5, 1.5, and 10 mg kg−1. The experimental group for each dose consisted of 2 mice per group. Both compounds were formulated in 0.5% HPMC containing 0.2% Tween 80 and were administered via oral gavage in an administration volume ranging from 200 to 300 μl, to maintain an equal compound dose relative to the weight of each mouse. Chloroquine was included in experiments as a positive control and dosed at 2.5, 5, and 10 mg kg−1. Chloroquine was formulated in saline solution. Additionally, two negative-control groups were included, one receiving 0.5% HPMC containing 0.2% Tween 80 only and the other containing saline only. Blood samples (2 μl) were collected on days 3, 4, 5, 6, and 7 to determine parasitemia levels by flow cytometry using a bidimensional staining technique, as previously described (23). Additionally, to determine the pharmacokinetic (PK) parameters, blood samples (10 μl) were collected via tail bleeding predose and at 0.25, 0.5, 1, 3, 5, 8, and 24 h postdose in 1.5-ml lithium heparin Eppendorf tubes to prevent blood coagulation. Samples were stored at −80°C prior to LC-MS/MS analysis of the sulfoxide and sulfone.
For the single-dose regimen of the sulfoxide, an experimental procedure similar to the one described for the quadruple-dose regimen was used, with the modification that each experimental group received a single dose of 0.1, 0.5, 1.5, 5, 15, and 50 mg kg−1 on day 3 following infection. Chloroquine was included in the experiment as a positive control and dosed at 50 mg kg−1. Blood samples (10 μl) were collected predose and at 0.25, 0.5, 1, 2, 4, 6, 8, 10, 24, 48, 72, and 96 h postdose for PK evaluation. The samples were stored at −80°C prior to LC-MS/MS analysis of the sulfoxide and its sulfone metabolite.
(ii) Dose-response relationship analysis. Data analysis was performed using nonlinear fitting to a logistic equation of log10(% parasitemia on day 7 after infection) compared to the untreated control group using GraphPad Prism 4 software. This analysis determines the pharmacodynamic readout in the form of ED90 and AUCED90 values.
The 90% effective dose (ED90) is defined as the dose in milligrams per kilogram that reduces parasitemia on day 7 following infection by 90% compared to untreated control group. AUCED90 is defined as the average estimated daily exposure required to reduce parasitemia from peripheral blood on day 7 following infection by 90% compared to the untreated control group. The equation used to fit the data is y = bottom + (top − bottom)/[1 + 10([logED50 − x] × Hill Slope)]. This equation is a four-parameter logistic equation where “bottom” is the y value at the bottom plateau and “top” is the y value at the top plateau. LogED50 is the x value when the response is halfway between the bottom and top. “Hill Slope” describes the steepness of the curve. This variable is also called the slope factor or the Hill coefficient. If it is positive, the curve increases as x increases. If it is negative, the curve decreases as x increases. A standard sigmoid dose-response curve (previous equation) has a Hill Slope of 1.0. When the Hill Slope is less than 1.0, the curve is shallower. When the Hill Slope is greater than 1.0, the curve is steeper.
The ED90 is calculated by fitting the variable y = log10(parasitemia at day 7 after infection) and the variable x = log10(dose level in mg kg−1), defined as an ordered pair for each individual of the study. The AUCED90 is calculated by fitting the variable y = log10(parasitemia at day 7 after infection) and the variable x = log10(AUC of compound during the first 24 h after the first drug administration, in μg h ml−1), defined as an ordered pair for each individual of the study. The ED90 and AUCED90 are calculated by interpolation of the x value that corresponds to antilog10(y = top − 1) in each respective best-fitted curve.
(iii) Pharmacokinetic analysis. The whole-blood concentrations of the sulfoxide and sulfone were determined using a quantitative LC-MS/MS method. Sample preparation was achieved with a protein precipitation extraction method, using 5 µl whole blood and 195 µl acetonitrile-methanol (ACN-MeOH) (80:20) for the NSG samples and 20 µl whole blood with 180 µl ACN-MeOH (80:20) for the C57BL/6 samples. After centrifugation, the supernatant was transferred to 96-well plates for analysis. LC-MS/MS was carried out on an AB Sciex QTrap 5500 instrument coupled to an Agilent 1260 high-performance liquid chromatograph (HPLC). Separation was achieved on a Poroshell EC-C18 column, using 0.1% formic acid in water as the aqueous mobile phase and acetonitrile as the organic mobile phase. The analytical limit of quantitation (LOQ) was 0.5 ng/ml. The combined accuracy and precision (% nom) statistics were ±15% for the standards and quality controls of both compounds. Concentration-versus-time data were used to derive the PK parameters using noncompartmental analysis on PK Solutions 2.0 (Summit Research Services, Montrose, CO, USA).
(iv) Metabolite analysis. Metabolites of the sulfoxide and sulfone were identified in healthy mice using PK samples as well as in mouse liver microsomes (see the supplemental material).
ACKNOWLEDGMENTS
We acknowledge the following institutions for their contribution to the completion of this study: the PK laboratory and the animal unit (specifically Trevor Finch) of the Division of Clinical Pharmacology, Department of Medicine at the University of Cape Town, where the animal work and analyses of samples were done; GSK, Tres Cantos, for training two staff members in the NSG mouse model for malaria and for doing the quadruple-dose regimen that enabled us to compare our data with that of an external laboratory.
We express our appreciation to the Medicines for Malaria Venture for the challenge grant (to L.W. and K.C.) that enabled us to implement and validate the NSG mouse model for malaria at the University of Cape Town. The University of Cape Town, South African Medical Research Council, and South African Research Chairs Initiative of the Department of Science and Technology, administered through the South African National Research Foundation, are gratefully acknowledged for support (to K.C.).
FOOTNOTES
- Received 8 February 2018.
- Returned for modification 29 May 2018.
- Accepted 7 September 2018.
- Accepted manuscript posted online 24 September 2018.
Supplemental material for this article may be found at https://doi.org/10.1128/AAC.00261-18.
- Copyright © 2018 American Society for Microbiology.
