Dose-fractionation studies of a Plasmodium phosphatidylinositol 4-kinase inhibitor in a humanized mouse model of malaria

ABSTRACT

UCT594 is a 2-aminopyrazine carboxylic acid Plasmodium phosphatidylinositol 4-kinase inhibitor with potent asexual blood-stage activity, the potential for interrupting transmission, as well as liver-stage activities. Herein, we investigated pharmacokinetic/pharmacodynamic (PK/PD) relationships relative to blood-stage activity toward predicting the human dose. Dose-fractionation studies were conducted in the Plasmodium falciparum NSG mouse model to determine the PK/PD indices of UCT594, using the in vivo minimum parasiticidal concentration as a threshold. UCT594 demonstrated concentration-dependent killing in the P. falciparum-infected NSG mouse model. Using this data and the preclinical pharmacokinetic data led to a low predicted human dose of <50 mg. This makes UCT594 an attractive potential antimalarial drug.

INTRODUCTION

Malaria is a life-threatening disease transmitted by female Anopheles mosquitoes infected with Plasmodium parasites. Despite important advances in understanding the disease, the number of malaria cases has only decreased globally from an estimated 251 million cases in 2010 to 249 million cases in 2022, and malaria still caused 608,000 deaths in 2022, of which 94% were within the African continent (1). Artemisinin combination therapies (ACTs), considered the best treatment for malaria, have recently exhibited delayed parasite clearance and reduced susceptibility/resistance in infected individuals in South-East Asia and Africa (2)^, (3). This emphasizes the urgent need for new antimalarial drugs with novel modes of action as well as new drug combination regimens to overcome parasite resistance to ACTs and other antimalarials.

We previously reported the 2-aminopyridine MMV390048, the first preclinical candidate inhibiting Plasmodium phosphatidylinositol 4-kinase (PI4K) to have entered human clinical trials where a single dose of 120 mg was likely to achieve cure in patients with Plasmodium falciparum malaria (4–6). MMV390048 was potent across the plasmodial life cycle except against hypnozoites and progressed to phase 2a clinical development (4, 5). However, further clinical development of the compound was halted due to adverse developmental toxicity signals, including diaphragmatic hernias and cardiovascular malformations in rats but not in rabbits, which were revealed in embryofetal studies (7). In light of this attrition, the identification of back-up candidates was desirable. Further structure-activity relationship exploration around MMV390048 led to the identification of UCT943 (8) and UCT594 (9). The development of UCT943 was halted during preclinical safety and toxicity assessment due to signals (unpublished data) unrelated to the development toxicity signals seen for MMV390048. UCT594, a carboxylic acid, was differentiated from a physiochemical property perspective compared to the neutral MMV390048 (Fig. 1) and the basic UCT943. UCT594 showed potent in vitro activity against multiple stages of the parasite life cycle, desirable pharmacokinetic (PK) properties in the rat, and remarkable in vivo efficacy in the Plasmodium berghei and Plasmodium falciparum NSG (NOD-scid IL-2R_Ƴ^null) mouse models of infection, with effective dose resulting in a 90% reduction in parasitemia (ED₉₀) values of 0.56 mg/kg and <0.10 mg/kg, respectively (9). In addition, when UCT594 was tested against a panel of resistant field isolates, the ratio of the maximum/minimum IC₅₀ (half maximal inhibitory concentration) values was 1.8-fold, suggesting a low risk for cross resistance with existing antimalarial drugs (9).

Fig 1.

Molecular structures of MMV390048 and UCT594.

To further assess the potential of UCT594 as a back-up compound to MMV390048 for preclinical development, cross-species in vitro and in vivo drug metabolism and pharmacokinetic (DMPK) profiling was carried out. Additionally, dose-fractionation studies in the Plasmodium falciparum NSG mouse model were performed to determine the driver of the in vivo efficacy. The DMPK profile and efficacy data together offer tools to better predict human PK parameters and the efficacious dose in humans.

RESULTS

In vitro efficacy profiling of UCT594

MMV390048 was identified to be an inhibitor of Plasmodium PI4K (4), a novel target expressed at all the life cycle stages of P. falciparum and Plasmodium vivax parasites (10). UCT594 has comparable activity to MMV390048 against P. falciparum (NF54) with an IC₅₀ of 32 nM. UCT594 inhibited the P. vivax PI4K enzyme with an IC₅₀ value of 24 nM, about sevenfold higher than the value seen for MMV390048 (3.4 nM). UCT594 was also assessed in competitive binding studies against an immobilized drug used to pull down PfPI4K from cell lysates, giving >80% inhibition at 10 µM (Table S1 in Supplementary information). This suggests that UCT594 also acts through inhibition of Plasmodium PI4K. UCT594 showed similar activity to MMV390048 against the human PI4Kβ ortholog (7.7 µM vs 1.0 µM for MMV390048), thus displaying a selectivity index over hPI4Kβ of >300- and >200-fold for PvPI4K and anti-Plasmodium activity, respectively (Table S1 in Supplementary information).

UCT594 primarily exhibited blood-stage activity against schizonts (Table S2 in Supplementary information), which correlates with a slow rate of kill seen in the in vitro parasite reduction ratio (PRR) assay (11) with a 48-h lag phase and log reduction ratios of 2.3 at 10-fold EC₅₀ (3D7 strain). The profile of UCT594 in the in vitro PRR assay was similar to that of MMV390048 (lag phase, 48 h; log PRR, 2.7) (4) and most closely matches that of atovaquone. Furthermore, the ex vivo susceptibility of UCT594 was tested against Plasmodium isolates collected from patients from Papua, Indonesia, a region endemic for multidrug-resistant strains of P. falciparum and P. vivax. The activity observed in these assays was comparable to that of MMV390048 (8) (Table S3 in Supplementary information). Interestingly, UCT594 displayed higher potency against P. vivax than against P. falciparum. When tested against other stages of the parasite life cycle, UCT594 was active against most stages of the parasite life cycle in vitro (Fig. 2).

Fig 2.

Antiplasmodium activity of UCT594 compared to that of MMV390048 against different stages of the parasite life cycle. Values indicate EC₅₀ values in nanomolar (nM); black, UCT594; brown, MMV390048; Pf, P. falciparum; Pc, Plasmodium cynomolgi; Pb, P. berghei; Pv, P. vivax; R/T/S, ring/trophozoites/schizont. (Reused with permission from MMV).

Besides blood-stage activity against the malaria parasite, UCT594 also showed and EC₅₀ = 125 nM in the standard membrane feeding assay (SMFA) indicative that it would reduce parasite transmissibility. Additionally, UCT594 was active against Plasmodium cynomolgi small and large forms with an EC₅₀ of 119 nM and 49 nM, respectively, when evaluated prophylactically. UCT594 was also prophylactically active against P. vivax small and large liver forms in vitro (EC₅₀’s <100 nM). It is not surprising that a steep decrease in activity was seen when compounds were assessed against Pc small and large forms in the relapse mode, which is in line with other PI4K inhibitors (4, 8, 12). Hence, UCT594 is not likely to provide a radical cure for malaria. UCT594 had sub-micromolar activity against early and late-stage gametocytes, but it only weakly inhibits the formation of male and female gametes (9).

In vitro safety profiling of UCT594

UCT594 was also evaluated in a CEREP panel of 98 assays (receptors, transporters, and ion channels). Included among the assays are the key ion channels hERG, Na_v1.5, Ca_v1.2, and K1.5. UCT594 did not inhibit the hERG current in the patch clamp electrophysiology assay at the highest concentration tested (IC₅₀ >33 µM), which suggests that there is a low risk of causing QT interval prolongation (a longer interval between the start of the Q wave and end of the T wave) on dosing in humans. The compound also had no activity against the Na_v1.5, Ca_v1.2, and K1.5 ion channels at the maximum concentration (33 µM) tested.

UCT594 was tested against a panel of four mammalian cell lines to evaluate cytotoxicity risks wherein the selectivity indexes over NF54 activity were greater than 1,000-fold in all cases, suggesting that myelosuppression in humans linked to cytotoxicity from the four mammalian cell lines is therefore unlikely to be a concern (Table S4 in Supplementary information).

Physicochemical characterization

Consistent with its nature as a weak carboxylic acid, UCT594 has a pK_a of 4–6 (Table 1). A more accurate value could not be determined due to its exceedingly low aqueous solubility at low pH, which necessitated the use of co-solvents. The compound demonstrated pH-dependent solubility and, as expected, was more soluble at pHs above its pK_a. Solid form stability was monitored by X-ray powder diffraction, and no crystal form changes were observed under the conditions used for solubility determination. UCT594 was also highly permeable in Caco-2 cells and had a negligible efflux (efflux ratio 1.4), predicting high in vivo absorption. Since solubility increases substantially at higher pHs as those in the upper intestine, it would be expected that bioavailability would also be high. Furthermore, the acid functionality in UCT594 resulted in a lower LogD_7.4 compared to MMV390048 containing the methylsulfonyl moiety.

TABLE 1.

Physicochemical properties of UCT594 and MMV390048

Property	UCT594	MMV390048
Mol wt (g/mol)	359.3	393.4
LogD (pH 7.4)	0.52	2.6
pKa	4–6	4.0
Thermodynamic solubility (µM)
pH 2.0	<0.64	1,881
pH 4.0	<0.64
pH 7.4	85	10
pH 10.0	26,719
SGF (pH 1.8)	16
FeSSIF (pH 5.0)	306	72
FaSSIF (pH 6.5)	50	37
Caco-2 A-B/B-A (106 cm.s−1)	41/59	39/23

In vitro metabolism

UCT594 was metabolically stable (below the limit of detection of parent) in microsomes across mouse, rat, dog, and human species (Table 2). To better measure hepatocyte stability, incubations were relayed over 20-h refreshing hepatocytes every 4 h. The full body clearance extrapolated from hepatocyte CL_int, f_u plasma and blood plasma ratio, was below 10% of liver blood flow across species.

TABLE 2.

In vitro metabolic stability of UCT594

Parameter	Human	Monkey	Dog	Rat	Mouse
Microsomal CLint,app (mL/min/kg)	< 10.4	–	<15.0	<20.9	<45.7
Hepatocyte CLint,app (mL/min/kg)	2.3	3.3	11	19	<71
Predicted in vivo CLb from hepatocytes (mL/min/kg)	0.32	1.7	1.8	2.5	<10
fu plasma	0.073	0.074	0.074	0.075	0.079
fu microsomes	0.71	–	0.80	0.77	0.78
Calc. fu heps	0.94	0.94	0.94	0.94	0.94
Blood/plasma ratio	0.9	–	1.0	0.9	1.1
Plasma stability (% rem. after 240 min)	95	–	97	100	101

^a -, not determined.

The major metabolite from the hepatocyte incubations was thought to be the acylglucuronide (Fig. 3) based on liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) analysis. The acylglucuronide retained some antiplasmodium activity, though it is unclear whether the activity is due to hydrolysis to UCT594 in plasmodium cell culture. The structure was confirmed by comparison of the high-performance liquid chromatography (HPLC) retention times of the found metabolite and synthetically prepared UCT594-AG (supplementary information) which were identical.

Fig 3.

Structure of UCT594 and its acylglucuronide metabolite UCT594-AG

In vivo pharmacokinetic studies

The pharmacokinetic profile of UCT594 was determined in mice, rats, dogs, and monkeys (Table 3). Concentrations of the acylglucuronide metabolite, UCT594-AG, were also determined during these studies. UCT594 was rapidly absorbed across species. Absorption was complete or almost complete (bioavailability >90%) in mice and monkeys, but lower in rats and dogs (63% and 54%, respectively). However, higher (>90%) bioavailability had previously been observed in rats using a different formulation (0.5% hydroxymethyl cellulose and 0.2% Tween80), showing that bioavailability can vary, depending on the formulation suggesting variable rates of solubilization (9). Consistent with in vitro data, the clearance of UCT594 was low across species, below 10% of the liver blood flow. The ratio of exposure of the acylglucuronide relative to that of the parent varied from 0.2 in mice to 5.2 in rats. It is consistent that rats, which had the highest levels of acylglucuronide, also had the highest clearance.

TABLE 3.

In vivo whole-blood pharmacokinetics of UCT594 in mice, rats, dogs, and monkeys^a

Parameter	Mouse	Rat	Dog	Monkey
PO Cmax (µM)	15.4	2.0	12.2	20.4
PO Tmax (h)	0.5	3.7	0.25	2.0
PO t½ (h)	4.1	6.7	36.6	6.8
IV Vd (L/kg)	4.2	10.9	1.7	2.9
IV Vss (L/kg)	2.4	7.8	1.0	1.9
IV CLb (mL/min/kg)	6.3	14.5	3.3	1.4
PO UCT594 AUC0-inf (min.µM)	13,996	1,245	4,576	19,483
PO UCT594-AG AUC0-inf (min.µM)	3,080	6,534	4,005	44,402
PO metabolite:parent AUC ratio	0.22	5.2	0.88	2.3
F (%)	≥100	63	54	95

^a PO, oral dose of 10 mg/kg; IV, intravenous dose of 2 mg/kg; C_max, maximum concentration; T_max, time required to each maximum concentration; t_1/2, elimination half-life; V_d, volume of distribution; V_ss, apparent volume of distribution at steady state; CL_b, whole-blood clearance; AUC_0-inf, area under the curve from time 0 extrapolated to infinite time; F, bioavailability.

Dose-fractionation studies of UCT594 in the NSG mouse model for malaria and in vivo pharmacokinetics

The dose-fractionation study in the NSG mouse with UCT594 assessed five oral doses ranging from 0.0625 to 5 mg/kg with each dose fractionated one-, two-, and fourfold. Hence, dosing was carried out as a single dose given over the 4 days of the study, two doses every 2 days, and four doses daily. The relevant whole-blood pharmacokinetic parameters for oral administration of UCT594 in the infected NSG mouse are presented in Table S5 (supplementary information). Concentrations of the glucuronide metabolite were determined in the 0.125, 0.0625, and 0.0125 mg/kg groups (Table S5). For these dosing groups, levels of the acylglucuronide in the NSG mice were very close to levels of the parent [area under the curve (AUC) ratios of 1]. Fractionation of the dose did not have a significant effect on exposure by AUCs, which were not statistically different (P > 0.05) comparing the single, two, or four doses (Table S5).

PK/PD relationship analysis of UCT594

In order to determine the in vivo EC₅₀ value of UCT594, the whole-blood concentration-time data and whole-blood parasite-time data were modeled sequentially using a PK/PD model adapted from the literature (8). The final PK/PD model was evaluated by visual inspection of the goodness-of-fit plots and visual predictive checks shown in Fig. S1 of the supplementary information. The PD model estimated parameters from total whole-blood concentrations are summarized in Table 4, while population PK parameters are shown in Table S6 of the supplementary information.

TABLE 4.

Estimated whole-blood PD parameters for UCT594 in NSG mice infected with P. falciparum^a

Parameter	Estimate	95% CI (lower, upper)
Kkill (1/h)	0.051	(0.048, 0.054)
EC50 (ng/mL)	5.28	(2.57, 10.9)
EC90 (ng/mL)	11	–
H	3 (fixed)	–
Kgro (1/h)	0.03 (fixed)	–

^a K_kill, parasite kill rate constant; EC₅₀, effective concentration required to produce the half maximal effect; EC₉₀, effective concentration required to produce the 90% maximal effect; H, Hill coefficient; K_gro, parasite growth rate constant.

The in vivo minimum parasiticidal concentration (MPC) was determined using equation 5 and then corrected for the difference in binding between the NSG model and human plasma to give a human-corrected in vivo MPC of 16.6 ng/mL (8).

The MPC of UCT594 was set as the threshold required for determining the PK/PD indices in the P. falciparum-infected NSG mouse model. The various PK/PD indices are listed in Table S7 in the supplementary information. Nonlinear regression analysis was performed for the total AUC_0-96h/MPC, C_max/MPC and percentage time above MPC against the parasitemia for each dose-fractionated treatment group 96 h post-treatment, as shown in Fig. 4. This indicated that AUC_0-96h/MPC, C_max/MPC and percentage time above MPC were highly predictive of parasitemia, making it difficult to discern the primary driver for activity. Of the three PK/PD indices, the AUC_0-96h/MPC and C_max/MPC correlated best (R² = 0.86). The percentage time above MPC correlation (R²) was 0.85 (Fig. 4). This indicates that UCT594 exhibited concentration-dependent killing in the P. falciparum-infected NSG mouse model. The total exposure level (AUC) that resulted in a 99% reduction in parasitemia was 0.481 µg.h/mL, which correlates well with previous work conducted (9).

Fig 4.

Relationship between the PK/PD indices AUC_0-96h/MPC (a), C_max/MPC (b), and %T > MPC (c) and the corresponding parasitemia of UCT594 after dose fractionation in P. falciparum-infected NSG mice. C_max, peak concentration; AUC_0-96, area under concentration-time curve from 0 to 96 h; R², coefficient of determination.

Prediction of human pharmacokinetics and efficacious dose

In vitro in vivo correlation was conducted using hepatocyte data, and it was found that the in vitro data underpredicted in vivo clearance by an average of sevenfold across mouse, rat, dog and monkey clearance data. Therefore, allometric scaling for a 70 kg healthy human based on maximum life span potential (MLP) interspecies scaling was used (13). In this method, human PK parameters were predicted using the preclinical PK parameters of unbound clearance (CL_u) and unbound volume of distribution at steady state (Vss_u) (Fig. S2). The PK parameters correlated well with the body weights across the species (R² = 0.91 and 0.97 for volume and clearance, respectively). Using these parameters, the predicted total human blood clearance was 0.7 mL/min/kg (Table 5), i.e., less than 10% of liver blood flow (14), indicative of a low clearance compound. The predicted total volume of distribution was 1.1 L/kg. The predicted mean residence time (MRT) was 25.5 h, and the predicted human half-life was 18 h.

TABLE 5.

Summary of predicted human PK parameters using MLP corrected allometry

Predicted human PK parameters
Vssu (L/kg)	16.3
Vsstot (L/kg)	1.1
CLu (mL/min/kg)	10.7
CLtot (mL/min/kg)	0.7
MRT (h)	25.5
T1/2 (h)	18

The human dose was calculated on the basis of the predicted human PK and the free efficacious AUC in the NSG mouse model. The AUC that resulted in a 99% reduction in parasitemia (0.481 µg.h/mL) and correcting for species binding differences between NSG mice (fu of 0.11) and human (fu of 0.07) gave an extrapolated human unbound AUC (fAUC = 0.757 µg.h/mL) (Equation 8). To calculate the single human dose, the fAUC was multiplied by the unbound clearance estimated by allometry and further divided by the average bioavailability observed across the species (Equation 9). Based on this, the calculated single dose was 42 mg (Fig. 5). Using the human PK parameters from allometry and the estimated dose, the human PK profile was simulated and resulted in a free C_max of 23.9 ng/mL and a free exposure of 757 ng.h/mL. This dose was well within the predicted dose range (10 mg–40 mg) using allometry and Wajima normalization and transformation of the monkey and dog PK data, and predicting the human dose required to maintain exposure above the MPC (16.6 ng/mL) for 7 days.

Fig 5.

Human PK projected from efficacious AUC in the P. falciparum mouse model for malaria with a single dose of 42 mg.

The predicted low exposure provides a good safety margin based on available in vitro toxicology data and will be re-evaluated once in vivo toxicology data become available. Additionally, the low exposure provides a margin for dose adjustment should it be required during first-in-human studies, and if deemed desirable toward staving off the development of resistance and/or allowing for variable human PK.

DISCUSSION

A limitation of MMV390048, the PfPI4K inhibitor that reached phase 2 clinical trials, is its low solubility (10 µM in pH 7.4 phosphate buffer) attributed, in part, to not having an ionizable functionality. The low solubility was probably an important contributor to the variability in its exposure observed in the phase I clinical study (5). By contrast the acid functionality of UCT594 imparts higher aqueous solubility at elevated pHs including that at the physiological pH of 7.4 (85 µM). The carboxylic acid also offers an avenue for salt selection that can be incorporated into formulation development for clinical use. Low dissolution rates in the intestine can be expected to increase the time to C_max and lower bioavailability, which can be exacerbated if Plasmodium-infected patients develop diarrhea or other intestinal disturbances (15). Hence, UCT594 may offer greater latitude from a drug pharmaceutics point of view, which will be important for successful progression through clinical trials.

The activity of UCT594 against asexual blood-stage parasites was comparable to that of MMV390048 (4). In addition, UCT594 showed potential as a prophylactic and chemoprotective agent, demonstrated by its impact on parasite liver-stage development. UCT594 had the lowest ED₉₀ from in vivo efficacy experiments tested in the NSG mouse model for any published compound of the class (4, 8). The ED₉₀ (<0.1 mg/kg) was significantly lower than that of MMV390048 (ED₉₀ 0.57 mg/kg)(16), previously reported UCT943 (ED₉₀ 0.25 mg/kg) (8), and aminopyrazine sulfoxide and sulfone (ED₉₀ 0.12 mg/kg for both) (17). That UCT594 showed superior in vivo efficacy in the NSG mouse model is likely due to its higher free exposure in the blood after equivalent oral dosing, the fAUC for UCT594 in the NSG mouse model for a 0.5 mg/kg oral dose was about 220 ng·h/mL while that for MMV390048 was only about 49 ng·h/mL.

Dose-fractionation studies are aimed at understanding whether efficacy of a given drug is driven by time or concentration above a defined threshold. This enables the selection of dosing regimens that optimize clinical efficacy while suppressing the emergence of resistant organisms (18). The majority of antimalarials in use today were developed empirically in humans prior to the implementation of detailed pharmacokinetics and PK/PD studies in animals. This has resulted in sub-therapeutic doses, especially in children and pregnant women (19, 20), the two most vulnerable populations when it comes to malaria. Delineating the efficacy driver turned out to be a challenge for UCT594 as it is extremely active in the NSG mouse model of malaria, with an ED₉₀ value of <0.1 mg/kg and an AUC_ED90 value of less than 0.2 µg.h/mL; therefore, there is a limitation due to the very low concentration of compound approaching the limit of quantitation that affords efficacy in the NSG mouse model. Another limitation of the in vivo model relates to the determination of the MPC wherein the dynamic range of parasitemia is lower (2–3 log reduction) than the 4–5 log reduction seen in the in vitro PRR assay. Both assays are limited by the ability to detect lower levels of parasite, but the PRR assay allows for parasite regrowth that extends the dynamic range.

Nonetheless, the results suggested concentration-dependent killing is more likely to afford a better clinical result than maximizing time of exposure. The significance is that the preference within the malaria drug discovery community is toward a single-dose oral drug that delivers a cure to ensure maximum compliance, which directs efforts to dose high enough to achieve coverage above the MPC for 7 days representing three life cycles of the parasite. However, higher doses run to the risk of encountering toxicity effects. Current ACTs are administered over a period of 3 days in rural and poor populations where adherence to the dosing regimen is difficult to monitor. Suboptimal dosing due to non-compliance is known to contribute to the development of resistance. Toward this end, human dose predictions for UCT594 were directed at meeting a single dose based on the efficacious AUC in P. falciparum-infected NSG mice. The single dose makes sense as it addresses the critical need for the clinic, but it may not be predicted to be optimal for efficacy by the dose-fractionation study in the NSG mouse. The preference would in fact be fractionation of the dose over multiple days. Doing so would also lower drug C_max, which is important for toxicity concerns as mentioned.

UCT594 also had a favorable pharmacokinetic profile across preclinical species evaluated with low clearances and moderate half-lives. The only metabolite detected across all the species was the acylglucuronide, UCT594-AG, which was formed at different levels and demonstrated differences in its rate of formation or rate of excretion. However, the levels of glucuronide in hepatocyte incubations were quite low and not predictive of the in vivo situation. The formation of the glucuronide is an important consideration in the selection of preclinical species for toxicology evaluations, as the expectation is that it would be formed in humans. As has recently been discussed, the presence of an acylglucuronide metabolite in circulation is not an inherent safety concern, as correlations with toxic effects have not been seen (21). In order to cover the level of acylglucuronide that is expected to be observed in humans, preclinical toxicology evaluations of UCT594 are planned to be conducted in rats and dogs as the levels of this metabolite observed in both species are sufficient to qualify expectations in humans.

The predicted human dose of 42 mg is comparable to that of previously reported compounds in the series, including MMV390048 (80 mg–100 mg) and UCT943 (50 mg–80 mg) (4, 8). This prediction might evolve as clinical data are collected in a similar manner as for MMV390048, where the optimized clinical dose was 120 mg based on simulations and PK/PD modeling of clinical data collected in the induced blood-stage malaria study (6). The predicted exposures for UCT594 achieve good safety margins over in vitro cytotoxicity and cardiac ion channel inhibition.

MATERIALS AND METHODS

Preparation of acylglucuronide

The acylglucuronide of UCT594 (UCT594-AG) was prepared in three steps from UCT594 and glucuronic acid (Supplementary information). The compound was used as a standard for Met ID studies in animal PK experiments.

In vitro profiling of UCT594

(a) Asexual blood-stage assays

UCT594 was evaluated against the adapted strain of P. falciparum 3D7 harvested from the NSG model of infection; this was done using the Plasmodium lactate dehydrogenase assay (22) The assay was performed in triplicate in two independent experiments.

(b) Ex vivo drug susceptibility assay against resistant P. falciparum and P. vivax clinical isolates from Papua, Indonesia

Drug susceptibility of P. falciparum and P. vivax isolates from Papua, Indonesia, was measured using a protocol modified from the WHO microtest as described previously (23, 24).

(c) In vitro log PRR

The rate of parasite killing by UCT594 was determined using limiting dilution experiments as previously described (11).

In vitro cytotoxicity

In vitro cytotoxicity of UCT594 was tested against L6 cells using the alamarBlue assay and against Vero and HepG2 cells, using an MTT [3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazoliumbromide] assay (25, 26).

Liver-stage assays

(a) P. cynomolgi liver-stage assay

Primary rhesus hepatocytes were infected in vitro with P. cynomolgi sporozoites, and the drug assays were performed as previously reported by Zeeman et al. (27).

(b) P. vivax liver-stage assay

A P. vivax liver-stage assay was implemented in human hepatocytes infected in vitro with P. vivax sporozoites, according to the protocol described in Sattabongkot et al. (28).

Gametocyte assays

In vitro gametocytocidal activity was determined using luciferase reporter lines specifically enabling screening against early-stage gametocytes (>90% stages I to III) and late-stage gametocytes (>95% stages IV and V) as per Reader et al. (29). Methylene blue (5 µM) and MMV390048 (5 µM) were routinely included as controls.

P. falciparum dual-gamete formation assay (DGFA)

Transmission-blocking activity of UCT594 was assessed in a DGFA, which utilizes a dual readout that individually and simultaneously reports on the functional viability of male and female mature-stage V gametocytes, as per Ruecker et al. (30).

SMFA

Transmission-blocking activity of UCT594 was measured in the SMFA using Anopheles stephensi mosquitoes (31).

Physicochemical and in vitro absorption, distribution, metabolism and excretion (ADME) assays

Physicochemical properties

The pK_a of the compound was determined by potentiometric titration using a Metrohm 809 Titrando autotitrator as described previously (16) The thermodynamic solubility of UCT594 was determined by equilibrating the solid material in different media at 37°C for 24 h. The buffers and solutions used were pH buffers (pH 2.0, 4.0, 6.0, 8.0, 10.0), simulated gastric fluid, fasted state simulated intestinal fluid, and Fed State Simulated Intestinal Fluid. The samples were then centrifuged, and the supernatant was analyzed using an Agilent 1100 HPLC with a Diode Array Detector. Chromatography was performed using a Waters XBridge C18 column (4.6 × 150 mm, 5 µm) using 0.1% tetrafluoroacetic acid in water and acetonitrile as the aqueous and organic mobile phases, respectively. Monitoring and quantitation were performed at a wavelength of 220 nm (reference wavelength 500 nm). The limit of detection was 0.23 µg/mL (0.64 µM). The residual solid in each solubility sample was examined by X-ray powder diffraction to check for stability of the solid form under the test conditions.

In vitro metabolism

The metabolic stability of UCT594 was evaluated in human, dog, rat, and mouse microsomes using an adaptation of a previously described method (32) Briefly, a 1 µM UCT594 (final dimethyl sulfoxide [DMSO] concentration 0.25%) was incubated with 0.4 mg/mL human (Xenotech, 50 individuals mixed gender pool of 50), dog (Xenotech, male, pool 6), rat (Xenotech, male Sprague-Dawley, pool 700), or mouse (male BALB/c, pool of 800) microsomes. The reaction was started by adding 1 mM NADPH and incubated for 60 min at 37°C, with samples taken at 0, 5, 10, 30, and 60 min. Each sample was quenched using ice-cold acetonitrile containing MMV394902 as internal standard. Midazolam, propranolol, and MMV390048 were used as controls and were treated similarly. The supernatant was analyzed by LC-MS/MS using an AB Sciex 4000 QTRAP coupled to an Agilent 1200 Rapid Resolution HPLC. Natural log-transformed percentage remaining values were plotted against time to give the degradation constant (k min⁻¹) as the gradient. Microsomal half-life and intrinsic clearance were calculated as described by Obach (32).

Hepatocyte stability was performed using cryopreserved hepatocytes in relay (33). Also, 1 µM UCT594 (final DMSO concentration 0.25%) was incubated with 0.5 × 10⁶ human, rat, dog, and monkey hepatocytes (BioReclamation IVT) in Williams E media. In addition, 5 × 4 h relays were performed for an effective incubation time of 20 h. Analysis was performed using an AB Sciex API 4000 mass spectrometer coupled to a Shimadzu SCL-30A series HPLC. The degradation constant (k min⁻¹) was calculated from the linear portion of the Ln % percent remaining vs time curve and used to calculate half-life and hepatocyte CL_int using established equations (34).

Binding to plasma, microsomal protein and albumax were determined by ultracentrifugation as previously described (33).

In vivo pharmacokinetic studies: experimental animals

Male Sprague-Dawley rats and male BALB/C mice were bred at the University of Cape Town Research Animal Facility, Cape Town, South Africa and were acclimatized at the Animal Unit located at the Division of Clinical Pharmacology, University of Cape Town, South Africa. Male beagle dogs and male cynomolgus monkeys were part of the established Experimental Sciences colony. Three animals were used for each experimental group (intravenous and oral administration). Mice were not fasted overnight and were allowed to eat ad libitum. Rats were fasted overnight prior to dosing, with food restored 5 h after drug administration. Dogs and monkeys were fasted overnight prior to dosing, with food returned ~4 h after drug administration; animals were permitted access ad libitum to water.

Formulation and administration

UCT594 was formulated for intravenous administration as a solution in 20% hydroxypropyl-β-cyclodextrin and administered as a slow bolus infusion to give a dose of 2 mg/kg. The oral dose was prepared as a suspension in PEG-400/PVP VA-64 and administered as a 10 mg/kg dose.

Sampling and processing

Blood samples were collected from the tail vein of mice and rats at different time points until 96 h post dose. They were collected in heparinized microcentrifugation tubes and immediately stored at −80°C pending analysis. These samples were extracted by protein precipitation using acetonitrile containing MMV394902 as internal standard. The mixture was then centrifuged, and the supernatant transferred to a 96-well plate for LC-MS/MS analysis. Calibration standards of UCT594 as well as its acylglucuronide, UCT594-AG, were prepared in blank mouse or rat blood and treated similarly.

Bioanalytical method

Quantitation of UCT594 and UCT594-AG was achieved by high-performance LC-MS/MS. Liquid chromatography was performed using an Agilent 1260 HPLC system (Sciex, Darmstadt, Germany), with the Agilent Poroshell (2.7 µm, 50 × 4.6 mm) column at an oven temperature of 40°C, coupled with an AB Sciex 5500 QTRAP mass spectrometer (Sciex, Darmstadt, Germany), equipped with a Turbo V ion source. The mobile phases used were A, 0.1% formic acid in water, and B, 0.1% formic acid in acetonitrile. A gradient was run with a flow rate 0.6 mL/min and a run time of 8 min. Sample preparation was achieved with a protein precipitation extraction method, using 10 µL whole blood and 100 µL ACN containing the internal standard (MMV394902). Data acquisition was performed using the Applied Biosystems software Analyst 1.6.3 (Sciex, Darmstadt, Germany). The calibration standards (0.5 ng/mL–6,250 ng/mL) and quality control standards were analyzed in triplicate for the parent compound (UCT594) and glucuronide metabolite (UCT594-AG). A quadratic regression, with peak area ratio (drug/internal standard) against concentration with 1/concentration (1/x) weighting, was fitted to the calibration curves. The combined accuracy (% nom) statistics were less than 15% for both standards and the quality controls of UCT594 and UCT594-AG.

Calculation of pharmacokinetic parameters

Mouse and rat pharmacokinetic parameters were calculated by non-compartmental analysis using PK solutions 2.0 (Summit Research Services, Montrose, CO, USA) with a method based on curve stripping. Dog and monkey PK parameters were calculated by non-compartmental analyses using WinNonlin Professional 5.2 (Pharsight Corporation, Mountain View, CA, USA) with a method based on multi-exponential curve fitting.

Pharmacokinetic parameters were estimated using stochastic approximation expectation maximization algorithm (35) that is implemented in Monolix 2019R2. Exploration of the raw data was performed using ggplot2 version 3.3.3, an R package (36). A one-compartment pharmacokinetic model with first-order absorption and linear elimination (Equation 1 – 3) was applied to fit the data

$${\displaystyle \frac{dAd}{dt}=-ka\times Ad}$$ (1)

$${\displaystyle \frac{dAc}{dt}=ka\times Ad-\left(\frac{CL}{Vd}\right)\times Ac}$$ (2)

$$Cc=Ac/Vd$$ (3)

The rate of change in parasitemia was modeled by the direct effect pharmacodynamic model (Equation 4).

$${\displaystyle \frac{dP}{dt}=Kgro\times P-Kkill\times \left(\frac{C{c}^H}{C{c}^H+IC{50}^H}\right)\times P}$$ (4)

The Kgro denotes parasite growth rate constant, P denotes parasitemia at time t hours, while Kkill is the first-order parasite kill rate constant. Cc denotes predicted concentration of UCT594 at time t hours from the pharmacokinetic model. IC₅₀ denotes the concentration needed to inhibit the growth of parasites by 50% while H represents the Hill coefficient of the steepness of the concentration-effect curve.

At MPC, the kill rate is 90% of the maximum kill rate. After using Kgro as observed in PfNSG (0.03 /h), and 0.051 /h as kill rate, the in vivo MPC was calculated according to equation 5.

$${\displaystyle MPC=EC50\times {\left(\frac{90}{100-90}\right)}^{1/H}}$$ (5)

The EC₅₀ denotes the effective concentration that results in 50% of the maximum response, while H represents the Hill coefficient of the steepness of the concentration-effect curve. The in vivo MPC was then used as the threshold to determine the indices responsible for observed efficacy.

PK/PD relationship analysis of UCT594 (dose-fractionation studies)

To assess the PK/PD indices responsible for observed efficacy, dose-fractionation studies were carried out in the humanized NSG mouse model for malaria ((37; (38).These experiments involve administration of an ascending series of doses of UCT594. The details of the target doses as well as formulations used are given in Table S7.

NSG mice engrafted with human erythrocytes (approximately 60%) were infected with 2 × 10⁷ P. falciparum Pf3D7^0087/N9 cells, a strain developed at GlaxoSmithKline (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. Two mice were used for each regimen, with the control mice receiving vehicle only. An additional control group was included that received 10 mg/kg chloroquine. Groups receiving the single dose were dosed on day 3, the groups receiving two doses were dosed 48 h apart (day 3 and day 5), and groups receiving four doses were dosed for 4 consecutive days, receiving each dose 24 h apart (day 3–day 6). The measured parasitemia (day 7) served as the PD parameter. Fresh samples of peripheral blood from P. falciparum-infected mice were stained with TER-119-PE (marker for murine erythrocytes) and SYTO-16 (nucleic acid dye) and then analyzed via flow cytometry as described previously (39)

Ascending oral PK of UCT594 were carried out to assess the dose proportionality and to determine the PK parameters which govern antimalarial efficacy which include the maximum concentration, the overall exposure, and the duration in which the drug concentration remains above the defined minimum therapeutic concentration or threshold concentration (40) The experimental data were evaluated in terms of drug concentration versus time. The C_max, AUC_0-96h, and AUC_0-t were estimated by non-compartmental analysis using PKanalix (PKanalix version 2019R2. Antony, France: Lixoft SAS). Nonlinear regression analyses of the log-transformed PK/PD indices and parasitemia were performed using GraphPad Prism 9 (San Diego, California USA) to determine the strength of the relationship and establish the main PK/PD index driving the in vivo efficacy.

Human dose prediction

Allometric scaling

Allometric scaling is based on the idea that physiological and anatomical properties of preclinical animals can be scaled across species with respect to their body weight (41) Allometry was applied to the parameters of volume of distribution at steady state (V_ss) and clearance from the preclinical species’ (mouse, rat, dog, and monkey) PK, which were plotted against body weight on a log-log scale. The allometric exponent (b) and coefficient (a) were derived by means of least-squares linear regression using a log (unbound clearance) CL_u) versus log (body weight, BW) plot by simple allometry (equation 6). When the exponent of the clearance vs body weight plot b in simple allometry is between 0.7 and 1, it has been observed that correcting for the MLP per species improves the accuracy of the prediction of the human CL_u (42). Thus, to improve the predictability of the human CL_u, MLP scaling was used (equation 7). Human CL_u was extrapolated using an assumed human body weight of 70 kg.

$${CL}_u=a\times {BW}^b\times f_u$$ (6)

where BW is the body weight, f_u is the free fraction in plasma, and a and b are the coefficient and exponent of the allometric equation, respectively.

$${CL}_u=a`\times {BW}^{b`}\times \frac{f_u}{MLP}$$ (7)

where BW is the body weight, f_u is the free fraction in plasma, MLP is the maximum life expectancy per species, and a’ and b’ are the coefficient and exponent of the allometric equation, respectively.

Calculation of efficacious human dose and simulation of human PK

The exposure level that resulted in a 99% reduction in parasitemia in NSG mice, observed as the main PK driver of the intended pharmacology, was used to calculate the unbound efficacious human AUC by means of equation 8. This corrected human free AUC was subsequently used to calculate the human dose required to reach the efficacious exposure, by means of equation 9.

$$fAUC,human=\frac{AUC,NSG\times fu,NSG}{fu,human}$$ (8)

where fAUC is the unbound efficacious AUC, f_u,NGS is the free fraction in NSG mouse plasma and f_u,human is the free fraction in human plasma

$$Dose=\frac{fAUCxCLu}{F}$$ (9)

where CL_u is the unbound human clearance from allometric scaling and F is the average bioavailability observed across the four preclinical species. Human PK was simulated by means of one-compartment analyses using Phoenix WinNonlin 8.4 (Certara LP, 210 North Tucker Boulevard Suite 350, St. Louis, MO 63101 USA).

Wajima transformation and simulation of human PK

The plasma concentration-time profile of UCT594 in humans was simulated by first normalizing the preclinical species’ concentration-time data using the mean residence time (MRT) and volume of distribution at steady state (V_ss). Thereafter, the normalized animal time points and concentrations were transformed into human concentration-time data using the allometric predicted human MRT and V_ss. Based on the assumption that the dog and monkey may represent humans more closely, the concentration-time profiles of dog and monkey were used to simulate the human PK profiles, and modeled to determine the single dose range (based on dog and monkey, respectively) needed to maintain concentrations of UCT594 above the modeled in vivo MPC (16.6 ng/mL) for ≥7 days.

ETHICS APPROVAL

All studies and procedures were conducted with prior approval of the animal ethics committee of the University of Cape Town (approval numbers 017/025, 017/026, and 016/010) in accordance with the South African National Standard (SANS 10386:008) for the Care and Use of Animals for Scientific Purposes (43), and guidelines from the Department of Health (44). The human biological samples were sourced ethically, and their research use was in accord with the terms of the informed consents.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aac.00842-24.

Supplemental material. aac.00842-24-s0001.docx.

Supplemental methods; Fig. S1 and S2; Tables S1 to S7.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.