Developmental Toxicity Study for the Antimalarial Drug Combination Cabamiquine‐Pyronaridine Using a Translational Approach

ABSTRACT

Background

The safety of pharmacological treatments during pregnancy is a significant concern, as many medications lack comprehensive reproductive safety data. This creates gaps in therapeutic options for pregnant women, particularly malaria, which poses serious risks to maternal and fetal health. Additionally, there is an urgent need to develop combination therapies for which reproductive safety data are currently unavailable, further complicating treatment options for this vulnerable population. The combination of cabamiquine and pyronaridine emerges as a promising solution; however, since no reproductive safety data exist for this combination, comprehensive evaluations are essential to ensure maternal and fetal well‐being.

Materials and Methods

The cabamiquine–pyronaridine combination was evaluated using an in vivo embryo‐fetal development (EFD) toxicity study in rats and a Zebrafish Embryo Developmental Toxicity Assay (ZEDTA), following regulatory frameworks. Zebrafish developmental toxicity data were further integrated using pharmacokinetic and toxicokinetic modeling to assess safety margins and predict potential toxicity at anticipated clinical study doses and exposures.

Results

Combined administration of cabamiquine and pyronaridine showed no significant impact on prenatal outcomes. Both compounds exhibited dose‐proportional increases in maternal systemic exposure, with no fetal mortality or adverse findings. No adverse effects on embryo‐fetal development were observed at the previously defined monotherapy NOAEL doses, indicating the absence of adverse drug–drug interactions. Safety assessments at higher doses in the ZEDTA revealed no toxicity effects at ongoing clinical study doses. Pharmacokinetics and toxicokinetics modeling do not suggest the potential of cabamiquine and pyronaridine to cause toxicity at the forecast clinical study doses and anticipated exposures.

Discussion

The combined findings from the rat EFD study, ZEDTA, and model‐based pharmacokinetic and toxicokinetic analyses consistently demonstrated the absence of developmental toxicity and adverse drug‐drug interactions across multiple complementary approaches. These data provide confidence that the combination is unlikely to pose developmental safety concerns at clinically relevant exposures.

Conclusion

These findings support the clinical development of the cabamiquine–pyronaridine combination for women of childbearing potential and provide a foundation for future clinical evaluation during pregnancy.

1. Introduction

Infectious diseases during pregnancy represent a major global health challenge, affecting millions of women each year and contributing substantially to maternal, fetal, and neonatal morbidity and mortality (Kourtis et al. 2014; Silasi et al. 2015). Physiological adaptations during pregnancy, including immunological modulation required to tolerate the semi‐allogeneic fetus, render pregnant women more susceptible to infectious diseases and may exacerbate disease severity (Mor and Cardenas 2010; Robinson and Klein 2012). As a result, infections acquired during gestation can result in adverse outcomes that extend beyond maternal illness to include impaired fetal development, preterm birth, and long‐term consequences for infant survival and health (Kumar et al. 2022). Among these conditions, malaria in pregnancy remains one of the most consequential and persistent threats, particularly in tropical and subtropical regions where Plasmodium transmission is endemic (Gore‐Langton et al. 2022; Reddy et al. 2023).

Malaria infection during pregnancy is recognized as major contributor to poor maternal health and associated perinatal outcomes (Berhe et al. 2023), reflecting both increased vulnerability to infection and pregnancy‐specific pathophysiological processes, such as placental sequestration of Plasmodium‐infected erythrocytes, which leads to inflammation, damages placental structures, and impairs nutrient and oxygen exchange (Ayres Pereira et al. 2016; Chua et al. 2021; Zakama et al. 2020). Clinically, malaria during pregnancy is associated with a broad spectrum of complications affecting both mother and child, including maternal anemia, obstetric complications, impaired fetal growth, preterm birth, low birth weight infants and increased risk of perinatal mortality (Guyatt and Snow 2004; Kapisi et al. 2017; Reddy et al. 2023; Wilcox 2001). Despite its well‐recognized impact, therapeutic and preventive options for pregnant women and women of childbearing potential (WOCBP) remain limited by safety considerations and declining effectiveness of existing regimens.

In the absence of an effective vaccine, the World Health Organization (WHO) provides essential guidelines for managing malaria during pregnancy, integrating both preventive and therapeutic measures (Table 1). Preventive strategies recommended by WHO include intermittent preventive treatment in pregnancy (IPTp) with sulfadoxine‐pyrimethamine (SP) and the use of insecticide‐treated nets (ITNs) to reduce exposure to mosquito bites (WHO 2025). However, the effectiveness of IPTp with SP is increasingly compromised by widespread resistance among malaria parasites, and no alternative preventive regimens have been approved for routine use (Peters et al. 2007; van Eijk et al. 2019). Additionally, SP is not advised during the first trimester due to potential teratogenic effects and limited safety evidence in humans (González et al. 2020; WHO 2014). Finally, in malaria‐endemic areas with high co‐infection rates, adverse drug–drug interactions raise safety concerns, particularly as IPTp‐SP is contraindicated for HIV‐infected women on cotrimoxazole prophylaxis (CTXp) due to increased sulfonamide‐related risks (Figueroa‐Romero et al. 2022; Peters et al. 2007). On the therapeutic side, WHO's recommendations have evolved over time. Initially, quinine combined with clindamycin was advised for treating first‐trimester infections due to concerns about the potential teratogenicity of artemisinin observed in animal studies (Clark et al. 2004; González et al. 2020). However, recent clinical evidence has prompted WHO to endorse artemisinin‐based combination therapies (ACTs), particularly artemether‐lumefantrine (Coartem), as safe and effective for treating uncomplicated P. falciparum malaria across all trimesters of pregnancy (Dellicour et al. 2017). Promising treatments like pyronaridine‐artesunate (Pyramax) are being explored (SAFIRE trial, NCT06962319), yet a significant gap in malaria prevention persists, as no fixed‐dose combinations are currently in development. While dihydroartemisinin‐piperaquine has been assessed as a potential preventive option, studies indicate that its effectiveness does not exceed that of SP, which remains associated with better birth outcomes (Eisenberg and Krieger 2023; Permala et al. 2017). This underscores the urgent need for new antimalarial therapies that are safe and effective for WOCBP and pregnant women.

TABLE 1.

World Health Organization (WHO) guidelines: Recommendations for the treatment of pregnancy‐associated malaria.

	Trimester
	1st	2nd	3rd
Prevention	—	Sulfadoxine–pyrimethamine	Sulfadoxine–pyrimethamine
Treatment (uncomplicated)	Quinine + clindamycin or coartem	ACTs	ACTs
Treatment (severe)	Artesunate (IV) + ACTs

Abbreviations: ACTs, artemisinin‐based combination therapies; Coartem, artemether‐lumefantrine; IV, intravenous.

To address this gap, cabamiquine (M5717, DDD107498), a potent Plasmodium elongation factor 2 inhibitor, has been shown to be well tolerated, non‐teratogenic, with fetuses exhibiting similar exposure to the drug as their dams (Gado et al. 2024). These findings support its development for use in pregnant women and WOCBP. Cabamiquine is proposed for a fixed‐dose combination with pyronaridine, already available combined with artesunate in Pyramax, that targets the heme detoxification, inhibits P‐glycoprotein (Pgp), and influences topoisomerase II (TOPO II) activity, leading to DNA damage in the parasite (Bailly 2021; Chu and Dorlo 2023). The rationale for the proposed cabamiquine and pyronaridine combination includes enhancing efficacy, reducing the likelihood of resistance development, and leveraging complementary pharmacokinetic properties that may support simplified dosing regimens (Maiga et al. 2024; Rottmann et al. 2020). Building on this, the combination has recently successfully completed Phase 2a clinical studies (NCT05689047 and NCT05974267).

From a safety perspective, the inclusion of pyronaridine introduces additional considerations, including the potential for increased adverse events (AEs) or alterations in drug exposure. Pyronaridine is known to accumulate in tissues and has been associated with dose‐dependent toxicities in repeat‐dose studies, with the liver identified as target organ (Chu and Dorlo 2023; Croft et al. 2012). Reproductive toxicity studies have not demonstrated teratogenic effects at maternal exposures relevant to human therapeutic use; retarded fetal growth and pregnancy loss have been observed at maternally toxic doses (Chu and Dorlo 2023; Croft et al. 2012). These findings together with the reported TOPO II inhibitory properties of pyronaridine dictated the need to reinforce the knowledge on the reproductive and developmental safety of the cabamiquine–pyronaridine combination beyond the data generated on the individual components alone. This is also in line with the regulatory frameworks for combination products, for example ICH M3 (R2), stipulating that based on the concerns deriving from the properties of individual components, nonclinical embryo‐fetal development studies with the combination are recommended to address potential hazard for humans. This is particularly relevant for antimalarial therapies, where treatment during pregnancy or inadvertent exposure in early gestation is common in endemic regions.

Building on our previous work with cabamiquine, we have evaluated the reproductive and developmental safety of the cabamiquine–pyronaridine combination using two complementary approaches (Figure 1): (i) in vivo embryo‐fetal development (EFD) toxicity study in rat, and (ii) a New Approach Methodology (NAM) employing the Zebrafish Embryo Developmental Toxicity Assay (ZEDTA). Mathematical modeling was employed to assess cross‐species safety margins and to determine whether the combination produced additive or synergistic effects on maternal or embryo‐fetal endpoints beyond those observed with individual components. Following ICH M3(R2) and S5(R3) guidance, these studies integrated existing safety data for each component to provide a comprehensive risk assessment supporting potential clinical use in pregnant women and WOCBP.

FIGURE 1.

Non‐clinical safety assessment support of cabamiquine‐pyronaridine combination use during pregnancy. EFD, embryo‐fetal development; NAM, New Approach Methodology.

2. Materials and Methods

2.1. Test Item

2.1.1. Embryo Fetal Development (EFD) Study

For the reported study in this manuscript, cabamiquine (free base, CAS 1469439–69‐7, batch 110,705–03) was supplied as a powder by Merck Healthcare KGaA, Darmstadt, Germany while pyronaridine tetraphosphate (CAS 76748–86‐2, batch PPNPGC014) was supplied as a powder by Shin Poong Pharmaceutical, South Korea. Cabamiquine and pyronaridine tetraphosphate were suspended under magnetic stirring in the vehicle 1 (0.25% Methocel K4M Premium aqueous solution in water) and vehicle 2 (0.5% (w/w) methylcellulose, low viscosity aqueous solution), respectively, to reach the target concentration. Based on stability data aliquots of cabamiquine were stored at room temperature (25°C ± 5°C) while aliquots of pyronaridine tetraphosphate were stored in a refrigerator (2°C–8°C) in amber glass containers (to reduce light exposure) to be dispensed on each dosing occasion. In the above‐mentioned conditions, formulations were shown to be stable for up to a week for cabamiquine (1–20 mg/mL) and up to 8 days for pyronaridine tetraphosphate (0.3–150 mg/mL). The homogeneity of the formulation used in each study was verified by visual inspection prior to use.

2.1.2. Zebrafish Embryo Developmental Toxicity Assay (ZEDTA)

Cabamiquine (free base, CAS 1469439‐69‐7) and pyronaridine tetraphosphate (CAS 76748‐86‐2) were supplied as a powder by Merck Healthcare KGaA, Darmstadt, Germany while warfarin (CAS 81‐81‐2) and saccharin (CAS 81‐07‐2) used as positive and negative controls were purchased from Sigma‐Aldrich, Darmstadt, Germany. Stock solutions were freshly prepared by dissolving compounds in DMSO. The stock solutions were diluted with an aqueous solution (H₂O, 294.0 mg/L CaCl₂ × 2H₂O, 123.3 mg/L MgSO₄ × 7H₂O, 64.8 mg/L NaHCO₃, 5.8 mg/L KCl, purchased from Merck KGaA, Darmstadt, Germany) to reach the final test concentration in 0.5% DMSO (v/v).

2.2. Ethics Statement

Non‐clinical EFD study was conducted at Labcorp Early Development Laboratories Inc., Madison, United States of America, in accordance with the most recent version of the United States Food and Drug Administration (FDA): Good Laboratory Practice (GLP) for Nonclinical Laboratory Studies, Code of Federal Regulations, Title 21 Part 58, and any applicable amendments. These studies were designed to fulfill the requirements of ICH S5(R3).

2.3. Embryo Fetal Development (EFD) Study in Rats

To evaluate the potential adverse effects of the cabamiquine‐pyronaridine combination on maternal and fetal health as well as fetal development, an EFD study was conducted using pregnant female rats. The compounds were administered sequentially via oral gavage, the same route intended for human administration, during the organogenesis period. To avoid accumulation, cabamiquine was administered orally every other day (EOD) (Gado et al. 2024; Hewitt et al. 2022) at 0, 24, or 80 mg/kg, and like previous studies, pyronaridine was administered daily at 0, 47, or 140 mg/kg from Gestation Day (GD) 6 through GD17, inclusive (Chu and Dorlo 2023; Croft et al. 2012). Maternal parameters measured included daily clinical observations, body weight (BW), body weight gains (BWG), and food consumption (FC). All surviving rats were euthanized on GD21, and maternal placental and gravid uterine weights were recorded, along with the number of corpora lutea, implantation sites and resorptions. All fetuses underwent external, visceral, and skeletal examinations for morphological alterations. Maternal blood samples were collected on GD6 (1st dose) and GD16 (6th dose for cabamiquine and 11th dose for pyronaridine) for toxicokinetics (TK) analysis. A description of study design can be found in Table 2.

TABLE 2.

Rat EFD study design summary.

Group	No. of animals: Toxicity	No. of animals: TK	Cabamiquine (mg/kg/occasion) a	Pyronaridine (mg/kg/day) b	Duration of treatment	TK
1	24	3	—	—	GD6 to 17	GD6 and GD16
2	24	6	80	—
3	24	6	—	140
4	24	6	24	47
5	24	6	80	140

Abbreviations: GD, gestation day; TK, toxicokinetic.

^a Doses expressed as the free base; dose volume 5 mL/kg. Route of administration = every other day (EOD) by oral gavage. Vehicle was 0.25% (w/w) Methocel K4M Premium aqueous solution.

^b Doses expressed as tetraphosphate; dose volume 5 mL/kg. Route of administration = daily by oral gavage. Vehicle was 0.5% (w/w) methylcellulose—Low viscosity aqueous solution.

2.4. Toxicokinetics

Blood samples were collected from three TK animals/group/time point in Groups 2 through 5 on GD6 and 16, both pre‐dose and at approximately 0.5, 1‐, 3‐, 6‐, and 24‐h post‐dose. Additionally, samples were obtained from the vehicle control group (Group 1) at the same time points, including an extra collection approximately 3 h post‐dose. Using NaF and K₂EDTA anticoagulant, the samples were diluted 1:1 with citric acid for TK assessment. Systemic concentrations of cabamiquine and pyronaridine were determined through validated bioanalytical methods by Labcorp‐Madison Bioanalytical involving protein precipitation and liquid chromatography–tan‐dem mass spectrometry (HPLC‐MS/MS) detection. Toxicokinetic parameters, including (C _max) and area under the plasma concentration‐time curve (AUC_0‐t) were derived using Phoenix WinNonlin version 8.1 software (Certara USA Inc., Princeton, New Jersey). The accumulation ratios for both cabamiquine and pyronaridine were calculated using the standard definition, that is, as: Accumulation ratio = (Cmax or AUC_0–24 on GD16)/(Cmax or AUC_0–24 on GD6).

2.5. Zebrafish Embryo Developmental Toxicity Assay (ZEDTA)

Wild‐type zebrafish (ZF, Danio rerio ) breeding and husbandry were conducted according to standard practices (Aleström et al. 2020), maintaining water temperature, conductivity, and pH within optimal ranges. All experiments were conducted under non‐GLP conditions. As previously described (Gado et al. 2024), at 2 h post‐fertilization (hpf), fertilized embryos were placed in 24‐well plates (10 embryos per well, 10 embryos per condition), embryo media was replaced (E3 medium buffered to pH 7.2–7.6 with 10 mM HEPES) and the corresponding compounds concentration added. Saccharin (300 μM) and warfarin (300 μM) were used as negative and positive controls, respectively. Plates were incubated at 26°C under a 12:12‐h light/dark cycle in a precision incubator (Memmert, Schwabach, Germany) for 5 days (120 hpf) without medium renewal. Lethality and morphological effects were assessed visually at 120 hpf using an inverted microscope. Larvae were considered dead in the absence of a heartbeat, tail detachment, somite formation, or in the presence of coagulation. Embryo morphology was performed based on the teratogenic endpoints described in Table S1.

2.5.1. Cabamiquine Bioavailability in Zebrafish Embryos and Test Water

The bioavailability of cabamiquine and pyronaridine tetraphosphate was assessed in exposed larvae to estimate fish uptake as previously described (Gado et al. 2024). At day 5 all the surviving larvae exposed to the same concentrations were pooled (in groups of 10 or less, depending on the number of dead embryos), washed with reconstituted water, transferred into Precellys lysing vials (Bertin Technologies, Montigny‐le‐Bretonneux, France) and frozen at −80°C until bioanalysis. Compound stability was evaluated by measuring drug concentrations in test water at 0 and 120 hpf. Samples were analyzed by liquid chromatography–tandem mass spectrometry using a Waters H‐Class UPLC system (Waters GmbH) coupled to a Qtrap API5500 mass spectrometer (SCIEX) with a Turbo Ion Spray source in positive ionization mode. Quantification was performed in Multiple Reaction Monitoring mode using Analyst software.

Embryos/larvae homogenate concentrations were used to estimate drug exposure per larva (ppm/larva). Internal effective concentrations (μM) were calculated as previously described (Achenbach et al. 2022; Gado et al. 2024):

$$\begin{array}{cc}\text{Internal effect concentration}\left[\mathrm{\mu M}\right]=\hfill & \left(\left(\left[\text{pmol}\mathrm{per}\text{embryo}/\text{larvae}\right]\mathrm{x}\left[{10}^{-12}\right]\right)\right.\hfill \\ \hfill & \left./\left(\left[\text{embryo}/\text{larvae volume in}\mathrm{L}\right]\times \left[{10}^{-9}\right]\right)\right)\hfill \\ \hfill & \times \left[{10}^{-6}\right]\hfill \end{array}$$

2.6. Modeling of Exposure‐Toxicity

The acquired data was used to develop an exposure toxicity model. For exposure, the nominal concentrations of the compounds, both individually and in combination, in the surrounding water were multiplied by an uptake factor (mean ratio of concentration per embryo/larvae and concentration in water, calculated for each studied concentration). The concentration time profiles (expressed as area‐under‐the‐curve between 0 and 120 h) within the zebrafish embryo/larvae were chosen as the driver of toxicity. For toxicity, the fraction of animals showing no abnormalities was modeled.

In the first step, the relationships for each drug alone were modeled using the sigmoidal maximum effect model:

with Toxicity (%) being the percentage of embryos/larvae for each studied concentration showing no abnormalities, Imax (%) being the maximum toxic effect, AUC (μM·h) representing the area under the drug concentration‐time profile in the embryo/larvae from 0 to 120 h, IC₅₀ (μM·h) the AUC displaying the half‐maximum effect and Hill (dimensionless) indicating the steepness of the concentration‐toxicity relationship.

In the second step, the combined drug effects were modeled. Therefore, the general pharmacodynamic interaction (GPDI) model was used (Wicha et al. 2017). The GPDI model quantifies a deviation from an additive toxicity, that is, synergy or antagonism, as a shift of the maximum effect (i.e., Imax) or the potency (i.e., IC₅₀). Also, the potency of perpetrator drug and the steepness of the interaction relationship is estimated.

Modeling was performed in R (Version 4.5.2, R Foundation for Statistical Computing) using the R package “gslnls” (Version 1.4.2).

2.7. Human and Pregnant Rat Pharmacokinetic Modeling

PK modeling and simulation were performed using Berkeley Madonna Version 10.6 (Marcoline et al. 2022) and the latest release of RStudio for data processing, model implementation, and visualization.

2.7.1. Human

PK simulations for cabamiquine and pyronaridine in humans were based on previously published population PK models. Human cabamiquine exposure from the whole blood concentration data was described using the model developed by (Courlet et al. 2023), consisting of a three‐compartment structure with linear elimination, a transit compartment absorption model combined with first order absorption, and an enterohepatic recirculation component implemented via a depot compartment releasing drug back into the absorption pathway. This model captures the complex PK of cabamiquine. Interindividual variability (IIV) was included on apparent clearance (CL/F), central volume of distribution (V2/F), bioavailability (F), absorption rate constant (ka), transit rate (ktr), mean transit time (MTT), and the central to depot transfer rate (k2g).

Human pyronaridine exposure was simulated using the population PK model as previously reported (Barber et al. 2024). This model is a two‐compartment model with first order absorption and first order elimination, developed from whole blood concentration data in healthy volunteers experimentally infected with P. falciparum. The model describes the rapid absorption (t _max 1–2 h) and long terminal half‐life (8–9 days) characteristic of pyronaridine. IIV was included on the absorption rate constant (ka), apparent volume of distribution (V), clearance (CL), and intercompartmental clearance (Q). Both published models were used without structural modification for human exposure simulations.

2.7.2. Rat EFD

PK models were developed separately for cabamiquine and pyronaridine using the rat TK data (Figures S2, S3, S4 and S5). For both compounds, a two‐compartment model with first order absorption and linear elimination adequately described the observed concentration–time profiles. IIV was only included for simulations on clearance (CL), bioavailability (F), and intercompartmental clearance (Q). Exposure metrics (C _max, AUC_0–24h) were compared between animals receiving both compounds and those receiving vehicle alone. Across the evaluated dose ranges, systemic exposure to both cabamiquine and pyronaridine was generally comparable between concomitant and vehicle treated groups, indicating no substantial PK interaction under the tested conditions. Parameter values for both rat PK models are provided in Tables S8 and S9.

2.8. Clinical Trial Simulations

To translate the findings into the clinical perspective, a PK‐Toxicodynamic (TD) model was developed. Therefore, the model from 2.6 was linked to population PK models from 2.7 for pyronaridine (Barber et al. 2024) and cabamiquine (Courlet et al. 2023) in humans. The AUC_0‐120h in each simulated concentration‐time profile was assumed to drive the toxicity. In a sensitivity analysis, C _max was also evaluated to drive toxicity. Doses of 180, 360, 540 and 720 mg for pyronaridine (as tetraphosphate) and 60, 100, 200, 330, 400, 600 and 660 mg for cabamiquine (as free base) were simulated (alone and combined) for 1000 virtual patients each. The percentage of virtual patients with no safety issues was calculated for each scenario.

3. Results

3.1. Combined Administration of Cabamiquine and Pyronaridine Shows no Significant Impact on Prenatal Outcomes

Rats were selected to assess the risks associated with prenatal exposure to cabamiquine and pyronaridine when administered together to dams. Previous monotherapy studies indicated similar maternal and EFD effects in both rats, supporting a single‐species approach. Dose selection was based on prior studies, identifying the NOAEL for EFD (NOAEL_EFD) as 80 mg/kg EOD for cabamiquine (Gado et al. 2024) and 140 mg/kg/day for pyronaridine (Chu and Dorlo 2023; Croft et al. 2012).

To establish a link between the observed effects and drug exposure, combined systemic exposure was assessed at GD6 and 16 (Table S2), revealing that both compounds exhibited dose‐proportional and similar exposure parameters (C _max and AUC_0‐24), regardless of administration as single agent or combined. Notably, both compounds exhibited higher concentration‐time profiles after repeated dosing, indicating accumulation in pregnant rats.

No mortality related to either drug was observed during the study. Of note, non‐adverse findings in all pyronaridine‐treated groups included discoloration of bedding, urine, and fur consistent with the formulation color. Administration of cabamiquine and pyronaridine, alone or in combination, resulted in significantly lower mean body weights (BW), body weight gains (BWG), and food consumption (FC) compared to the control group (Table 3), with the most pronounced effects observed in animals receiving 80 mg/kg/occasion cabamiquine (Group 2) and the combination 80/140 mg/kg/occasion (cabamiquine/pyronaridine; Group 5), where statistically significant differences in mean BW were observed starting from GD12 and persisted through GD21. Throughout the dosing period mean BWG was reduced by −18.5%, −13.8%, and −35.4% for Groups 2, 3, and 5, respectively, compared to controls. Furthermore, Group 5 displayed significantly lower mean BW when compared to Group 3 (140 mg/kg/day pyronaridine) during the post‐dosing interval (GD18–21). Additionally, Group 5 had lower mean BWG when compared to Group 2 during the overall dosing (GD6‐18) and post‐dosing intervals (GD18‐21). Reductions in BWG were accompanied by decreases in carcass weight (Table 4) (10.0% Group 2, −6.7% Group 3, −5.7% Group 4 and −14.2% Group 5) and fetal weight at necropsy, particularly in Groups 2 and 5. Consistent with the observed BW and BWG changes, mean FC was significantly reduced in all treatment groups relative to control throughout the dosing period, with the greatest reductions occurring in the higher‐dose groups (20.9% Group 2, −19.8% Group 3 and −30.8% Group 5), and after cessation of dosing, FC remained significantly lower in all groups except Group 4 (24/47 mg/kg/occasion), which maintained FC values comparable to controls and higher than those observed in the monotherapy groups (Group 2 and 3). In contrast, animals in Group 4 exhibited BW and BWG comparable to controls throughout gestation, suggesting that the lower‐dose combination mitigated the adverse effects on maternal nutritional status seen at higher doses.

TABLE 3.

Summary of cabamiquine and pyronaridine‐related changes in body weight gain (BWG), corrected body weight and food consumption (FC) (mean value in grams, percent of difference compared with the control in bracket (%)).

	Group 1	Group 2	Group 3	Group 4	Group 5
No. of pregnant Dams	22	24	23 a	21	24
Maternal BWG (g)
GD6‐16	65	53*** (−18.5%)	56** (−13.8%)	62 (−4.6%)	42*** (−35.4%)
GD6‐21	77	57*** (−26.0%)	67* (−13.0%)	71 (−7.8%)	46 a *** (−40.3%)
GD18‐21	12	4* (−66.7%)	11 (−8.3%)	9 (−25.0%)	4 a * (−66.7%)
Maternal FC (g/day)
GD6‐18	253	200*** (−20.9%)	203*** (−19.8%)	224*** (−11.5%)	175*** (−30.8%)
GD6‐21	302	235*** (−22.2%)	245*** (−18.9%)	268** (−11.3%)	211*** (−30.1%)
GD6‐18	49	35*** (−28.6%)	42* (−14.3%)	44 (−10.2%)	36*** (−26.5%)

Note: *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. % difference versus controls; corrected body weight = terminal body weight—Gravid uterus weight, Group 1: control; Group 2: 80 mg/kg cabamiquine; Group 3: 140 mg/kg pyronaridine; Group 4: 24 mg/kg cabamiquine + 47 mg/kg pyronaridine; Group 5: 80 mg/kg cabamiquine + 140 mg/kg pyronaridine.

^a One animal was sacrificed in a moribund condition due to a dosing error.

TABLE 4.

Summary of cabamiquine and pyronaridine‐related changes in cesarean section parameters (mean value in grams, percent of difference compared with the control in bracket (%)).

	Group 1	Group 2	Group 3	Group 4	Group 5
No. Examined	22	21	23	21	22
Mean carcass weight (g)	218.39	196.59** (−10.0%)	203.77* (−6.7%)	205.98* (−5.7%)	187.33** (−14.2%)
Mean fetal weight (g)	5.094	4.649 (−8.7%)	4.763 (−6.5)	4.943 (−3.0%)	4.551 (−10.7%)

Note: *P ≤ 0.05; **p ≤ 0.001. % difference versus controls. Group 1: Control; Group 2: 80 mg/kg cabamiquine; Group 3: 140 mg/kg pyronaridine; Group 4: 24 mg/kg cabamiquine + 47 mg/kg pyronaridine; Group 5: 80 mg/kg cabamiquine + 140 mg/kg pyronaridine.

Pregnancy rates were 100% for groups receiving 80/140 mg/kg/occasion, 80/0 mg/kg/occasion, and 0/140 mg/kg/occasion, 87.5% for 24/47 mg/kg/occasion, and 91.7% for 0/0 mg/kg/occasion, with implantation occurring between GD4 and 5, prior to dosing, indicating that any lack of pregnancy was unrelated to the vehicle control or test articles. Groups 2 and 5 exhibited reduced mean fetal weights (−8.7% and −10.7%, respectively) relative to controls (Table 4), which were considered secondary to treatment‐related reductions in maternal BW. This interpretation is supported by the parallel between maternal BWG reduction and fetal weight decrease across groups (Tables 3 and 4), and notably by Group 4, which showed neither significant maternal BWG reduction (−7.8%) nor meaningful fetal weight decrease (−3.0%), consistent with the absence of a primary fetal toxicity effect. The absence of fetal weight reduction in the low‐dose combination group, in the context of preserved maternal nutritional status, provides strong evidence that the fetal weight changes observed in Groups 2 and 5 are secondary to maternal toxicity rather than a direct embryo‐fetal effect of the compounds. However, other parameters, including the number of live fetuses, dead fetuses, resorptions, and corpora lutea, showed no significant differences between the dosing groups and the control group (Table S3).

External, visceral, and skeletal examination of the fetuses showed no abnormalities associated with cabamiquine or pyronaridine (Table S4). Although fused mandibular/zygomatic arch was observed in Groups 2 (1 fetus; 2.38% fetal, 5% litter incidence), Group 3 (4 fetuses; 3.55% fetal, 13% litter incidence), and Group 4 (3 fetuses; 2.82% fetal, 14% litter incidence), this finding was considered incidental and not test article‐related. Fetal incidences across all groups were within the Labcorp historical control range (fetal incidence 3.61%), and the complete absence of this finding in Group 5, which received equivalent or higher doses of both compounds, is inconsistent with a treatment‐related effect. Overall, the combined administration did not adversely affect EFD in pregnant rats, and the NOAEL for maternal toxicity was established at 24 mg/kg cabamiquine co‐administered with 47 mg/kg pyronaridine while the NOAEL_EFD consistent with the highest tested dose of 80 mg/kg cabamiquine free base co‐administered with 140 mg/kg pyronaridine tetraphosphate.

3.2. Zebrafish Embryo Developmental Toxicity Outcomes Through Exposure‐Toxicity Modeling Indicates no Toxicity at Clinically Relevant Exposure

To support the rat EFD study outcomes and in alignment with the evolving regulatory framework for alternative methods in developmental toxicity testing (Powles‐Glover et al. 2025), we conducted the zebrafish embryo developmental toxicity assay (ZEDTA) on pyronaridine and the cabamiquine–pyronaridine combination (Tables S5, S6, and S7). Cabamiquine was not included as a single agent in this series; however, previous work found no developmental malformations in zebrafish embryos up to 100 μM (Gado et al. 2024). Prior to evaluating the combination, we therefore characterized pyronaridine's effects and uptake in embryos and larvae. Consistent with Truong et al. (2025), pyronaridine produced concentration‐dependent toxicity in zebrafish with nominal concentrations concentrations below 3 pmol were well tolerated, whereas higher doses caused increased mortality and developmental abnormalities (Truong et al. 2025). A 5‐day uptake‐kinetics study with pyronaridine, conducted as in our cabamiquine work, showed substantial accumulation, with embryo/larvae concentrations increasing from 0.6 pmol at 2 hpf to 7.14 pmol at 120 hpf, representing an approximately 12‐fold rise after a nominal 1 μM exposure (Table S7). Combination TD interactions were evaluated by integrating internal exposure data with scored morphological abnormalities and lethality using a mathematical exposure–response model to derive exposure–toxicity relationships for each agent as monotherapy and in combination, with the resulting dose–response curves and interaction effects summarized in Figure S1.

Compound effects on embryos were analyzed using AUC_120h, representing embryo/larvae exposure over the full 120 hpf study period, to provide a comparable metric of pyronaridine and cabamiquine exposure for cross‐species comparison with rat and human data. AUC_120h was estimated conservatively by assuming the final measured concentration remained constant for the remainder of the 120 hpf interval and integrating concentration over time. Cabamiquine monotherapy showed no statistically significant exposure–toxicity relationship whereas for pyronaridine, a significant relationship between exposure and the fraction of zebrafish embryos/larvae showing no toxicity effect was detected, with an IC₅₀ of (95% CI) of 5396 μM·h (4408–6385 μM·h) with a Hill factor of 3.4 (1.6–5.3). The inhibitory effect model (Imax) was fixed at 100% as no toxicity effect approached 0% at the highest pyronaridine exposure. In combination, increasing cabamiquine concentrations led to greater pyronaridine‐associated toxicity that was accompanied by increased pyronaridine uptake by the larvae (Table S5). Specifically, pyronaridine uptake in the 100 μM cabamiquine + 1 μM pyronaridine group was 4.7‐fold higher than in the 1 μM cabamiquine + 1 μM pyronaridine group, consistent with cabamiquine‐dependent potentiation of pyronaridine exposure and resultant toxicity. The higher toxicity can be directly observed in the heatmap (Figure 2) in the uppermost row (cabamiquine AUC_{0‐120 h} of 72,332 μM·h), where lower pyronaridine exposure induced toxicity. In line with that, the GPDI model quantified that the IC₅₀ of pyronaridine toxicity was reduced from 5396 to 1565 μM·h, if cabamiquine exceeded 61,100 μM·h (54,760–67,442 μM·h, potency of the interaction with a Hill factor of 20).

FIGURE 2.

Heatmap detailing the fraction of animals showing no developmental abnormalities (color gradient) in relation to mono (outer rows) and combine exposure (inner area) expressed as AUC_{0‐120 h} (μM·h).

Finally, the zebrafish embryo developmental toxicity NOAEL was established at 293 μM (corresponding to AUC_{0‐120 h} of 35,121 μM·h) cabamiquine free base co‐administered with 13 μM (corresponding to AUC_{0‐120 h} of 1528 μM·h) pyronaridine tetraphosphate.

3.3. Overall Exposure in Human Is Lower Than in Rat EDF and ZEDTA Studies

The comparative unbound concentration–time profiles of cabamiquine and pyronaridine across species (Figure 3) demonstrate that systemic exposure in humans remains notably lower than the exposures achieved in the rat EFD studies and in the ZEDTA.

FIGURE 3.

Comparative unbound plasma concentrations of cabamiquine (as free base) and pyronaridine (as tetraphosphate) in humans, rats, and zebrafish embryos. Preclinical species reach substantially higher exposures than humans. Orange lines = mean concentration profiles in rats (24/80 mg/kg cabamiquine; 47/140 mg/kg pyronaridine), black lines = human simulations (Barber et al. 2024; Courlet et al. 2023), and green line = zebrafish embryo uptake NOAEL (cabamiquine = 293 μM and pyronaridine = 13 μM); Orange area = IIV (30%) on clearance (CL), bioavailability (F), and intercompartmental clearance (Q) and black area IIV based on (Courlet et al. 2023; Barber et al. 2024); fu = unbound fraction (0.173 human and zebrafish embryo, 0.189 rats for cabamiquine; 0.05 human and zebrafish embryo, 0.08 rats for pyronaridine).

Each panel depicts the unbound fraction of the respective compound in plasma or embryo homogenate over a 300‐h post‐dose interval, to enable direct comparison across species. Human profiles (Figure 3, black curves) reflect the characteristic PK behavior of both compounds, rapid absorption followed by a prolonged terminal elimination phase, yet unbound concentrations remain consistently and markedly below those observed in preclinical species (Figure 3, orange curves = rats and green line = zebrafish embryos) at doses used to establish developmental safety margins.

For cabamiquine, unbound human plasma concentrations following 330 mg and 660 mg single oral doses remain well below 1 μM throughout the observation window, consistent with the compound's long elimination half‐life (146–193 h), extensive distribution, and high degree of plasma protein binding (fu = 0.173, R _b  = 1.3). In contrast, pregnant rats administered with 24 or 80 mg/kg EOD exhibited substantially higher systemic exposure, with C _max values of 688–2288 ng/mL and AUC_0–24 values of 9900–33,100 h·ng/mL (Table 5). These exposures increased further due to measurable accumulation (accumulation ratios up to 2.49). Overall, the rat exposures were approximately 5‐ to 10‐fold higher systemic exposure than anticipated at the efficacious human dose.

TABLE 5.

Simulated mean cabamiquine exposure in humans and pregnant rats of GD16, enabling a direct cross‐species comparison of systemic exposure under clinically relevant and toxicokinetic dosing conditions.

	Human		Rat
	330 mg	660 mg	24 mg/kg	80 mg/kg
Cmax,fu [μM]	0.055	0.11	0.28	0.943
AUC0‐24h,fu [μM*h]	0.97	1.95	4.05	13.53
Cmax [ng/ml]	149	299	688.16	2288.2
AUC0‐24h [μg/mL*h]	2.6	5.2	9.9	33.1
AR Cmax				1.26–1.82
AR AUC0‐24h				1.78–2.49
t1/2, eff				~48 h

Note: AR , accumulation ratio based on NCA; AUC_0‐24h, area under the curve for time 0–24 h; C _max, maximum mean concentration; fu , unbound fraction; t_1/2, _eff, effective half‐live.

Pyronaridine exhibited a similar interspecies exposure differential. Human unbound concentrations following 360 mg and 720 mg doses remained well below those achieved in rats at 47 or 140 mg/kg/day, where C _max values reached 859–2556 ng/mL and AUC_0–24 values increased up to 39,400 h·ng/mL (Table 6). Accumulation ratios for pyronaridine were higher than for cabamiquine (up to 3.66 for AUC_0–24), reflecting its long half‐life, extensive tissue distribution, and known propensity for intracellular sequestration.

TABLE 6.

Simulated mean pyronaridine exposure in humans and pregnant rats of GD16, enabling a direct cross‐species comparison of systemic exposure under clinically relevant and toxicokinetic dosing conditions.

	Human		Rat
	360 mg	720 mg	47 mg/kg	140 mg/kg
C max,fu [μM]	0.01	0.02	0.075	0.224
AUC0‐24h,fu [μM*h]	0.093	0.186	1.15	3.45
C max [ng/ml]	191	396	859	2556
AUC0‐24h [μg/mL*h]	1.7	3.38	13.23	39.4
AR Cmax				1.64–2.88
AR AUC0‐24h				2.36–3.66
t1/2, eff				~51 h

Note: AR, accumulation ratio based on NCA; AUC_0‐24h, area under the curve for time 0–24 h; Cmax, maximum mean concentration; fu , unbound fraction; t_1/2, _eff , effective half‐live.

Zebrafish embryos reached internal concentrations of cabamiquine and pyronaridine that exceeded the nominal aqueous concentrations by more than one order of magnitude (Table S5), indicating efficient compound uptake and sustained systemic exposure throughout the organogenesis window. Across species, the exposure profiles presented in Figure 3 and the toxicokinetic parameters listed in Tables 5 and 6 show that systemic exposures in dams and zebrafish embryos were substantially higher than those measured in humans at clinically relevant doses. In both preclinical models, exposure multiples frequently surpassed 5‐ to 10‐fold relative to human therapeutic levels, and no developmental toxicity was observed under these conditions. These data define wide developmental exposure margins for cabamiquine and pyronaridine and establish the systemic exposure ranges achieved in each model.

3.4. Model‐Based Simulations Support the Safety Use of Cabamiquine and Pyronaridine During Pregnancy

We used the conceptual approach of a previously published PK–PD framework to construct a PK–Tox model incorporating zebrafish‐derived developmental toxicity data. The model characterizes exposure and developmental safety to define acceptable dose ranges for WOCBP and pregnant women, using the same clinical dose levels as in the prior efficacy predictions (Maiga et al. 2024). The model translates the preclinical toxicity studies in zebrafish larvae/embryos into clinical perspective by coupling the toxicity parameters with clinically relevant PK profiles. Predicted toxicity, expressed as the fraction of virtual patients (n = 1000 per dosing regimen) displaying no toxicity effect is shown in Figure 4. The model used total concentrations for both humans and zebrafish larvae, as the uptake concentrations determined in the larvae represent total concentrations and we assumed that the fraction unbound in humans and zebrafish is similar for both cabamiquine and pyronaridine. Hence, also placenta crossing (which is driven by the unbound drug) is intrinsically considered in our model.

FIGURE 4.

Clinical trial simulations for exposure‐toxicity. Distribution of the outcome of ‘no abnormalities’ of the clinical trial simulations linking the AUC_0‐120h (simulated from population PK models of cabamiquine and pyronaridine in healthy volunteers) to the zebrafish exposure‐toxicity model. Median (black horizontal line), n = 1000 simulations for each dosing scenario.

The model predicted no toxicity effect when using the AUC_0‐120h as driver of the toxic effect in all investigated dosing scenarios, even when the quantified interaction (inactive cabamiquine potentiated the toxicity of pyronaridine in the zebrafish model) was considered in the simulations demonstrating the wide safety margin. The simulations using C _max led to identical results.

4. Discussion

The safety of pharmacological treatments during pregnancy is a critical concern, as many medications lack comprehensive reproductive safety data. This results in significant gaps in therapeutic options for pregnant women, particularly regarding malaria, which poses serious risks to maternal and fetal health.

Each year, up to 120 million pregnancies are at risk of malaria, resulting in about 819,000 infants born with low birth weight and contributing to 75,000–200,000 infant deaths (Steketee et al. 2001). Pregnant women are highly vulnerable because of the pregnancy‐associated alterations in cell‐mediated and humoral immunity, while the placenta affords a permissive microenvironment for Plasmodium sequestration (Ayres Pereira et al. 2016; Chua et al. 2021; Zakama et al. 2020). The resulting impaired development that can extend well beyond birth and produce lasting consequences, including increased susceptibility to infections (i.e., malaria, measles, Streptococcus pneumoniae and tetanus) (Boudová et al. 2017; Rent et al. 2023), delayed growth (Minwuyelet et al. 2025), and persistent neurodevelopmental and behavioral deficits (Buchwald et al. 2022; Weckman et al. 2021) that may continue into childhood and adulthood.

Current WHO‐recommended prevention therapy, IPTp‐SP, reduces P. falciparum morbidity (WHO 2024) but has uncertain efficacy against P. vivax (Minwuyelet et al. 2025; Nosten et al. 1999), which also contributes to maternal anemia and adverse pregnancy outcomes, and is problematic in certain co‐infection settings such as HIV (Figueroa‐Romero et al. 2022; Peters et al. 2007). Candidate regimens such as dihydroartemisinin–piperaquine have advanced into clinical evaluation; however, outcomes to date have been mixed and warrant further assessment (Eisenberg and Krieger 2023; Permala et al. 2017).

There is an urgent need for preventive antimalarials that combine proven efficacy with a well‐characterized reproductive safety profile for WOCBP and pregnant women. Ideal IPTp candidates should have an acceptable reproductive‐toxicity profile, a long half‐life allowing for monthly dosing and single‐dose administration (Figueroa‐Romero et al. 2022).

In response to this need, we evaluated the embryo‐fetal safety of a cabamiquine–pyronaridine combination in rats. Dose selection for the EFD study was guided by prior developmental and reproductive toxicity (DART) data for each agent as monotherapy. Earlier EFD investigations documented placental transfer of both compounds, with cabamiquine attaining fetal exposures comparable to maternal concentrations, whereas fetal pyronaridine levels were lower (4%–7%) than those in dams (Table S10). Although neither compound was teratogenic as a monotherapy, it was crucial to assess whether co‐administration could result in drug‐drug interactions affecting embryo‐fetal development. At the selected NOAELs (cabamiquine 80 mg/kg; pyronaridine 140 mg/kg), the combination did not result in any treatment‐related adverse embryo‐fetal findings and no evidence of teratogenicity. Furthermore, the systemic exposures at these NOAELs yielded favorable safety margins (Table 7) (14/7‐fold for cabamiquine and 37/18‐fold for pyronaridine) versus projected human exposures, supporting further consideration of the combination in pregnancy.

TABLE 7.

Safety margins derived from nonclinical studies in rats and zebrafish relative to projected human exposures.

Compound	Model	AUC0‐24 (μM·h)	Safety margins vs. ZF	Safety margins vs. Rat
Cabamiquine	Zebrafish	7024.3
	Rat	13.53
	Human (330 mg)	0.97	7241	13
	Human (660 mg)	1.95	3602	7
Pyronaridine	Zebrafish	305.64
	Rat	3.45
	Human (360 mg)	0.093	3286	37
	Human (720 mg)	0.186	1643	18

Building on data from the rat EFD study, we further explored potential toxic interactions using a novel approach by coupling ZEDTA with mathematical modeling to assess TD interactions. This integration allows for a comprehensive evaluation of TD interactions across various concentrations, recognizing that drug–drug interactions can differ considerably depending on exposure. Adapting a previously developed framework that linked in vitro efficacy data with modeling to identify pharmacodynamic (PD) interactions and inform PKPD dosing, we repurposed this approach to assess embryotoxicity rather than efficacy. ZEDTA complements this strategy by allowing testing at concentration ranges beyond those achievable in traditional animal studies (where maternal toxicity limits exposure), enabling direct embryo exposure, offering a rapid, cost‐effective new approach methodology (NAM) that aligns with the 3Rs principle and the 3rd revision of the ICH S5 guideline (Powles‐Glover et al. 2025).

Zebrafish embryos achieved internal cabamiquine and pyronaridine concentrations exceeding nominal aqueous concentrations by more than one order of magnitude, confirming efficient uptake and sustained exposure throughout the entire organogenesis window (Table S5). In this model, cabamiquine remained non‐toxic even at higher concentrations, consistent with prior published data (Gado et al. 2024). By contrast, pyronaridine produced dose‐dependent toxicity in zebrafish, consistent with previous results (Truong et al. 2025), with effects concentrated at estimated zebrafish uptake concentrations of 29–38 μM. A plausible explanation for the observed zebrafish toxicity is that pyronaridine, in addition to inhibiting haemozoin formation, inhibits TOPO II, a key enzyme in DNA replication, recombination, and repair. The effect window (≈29–38 μM) closely overlaps the reported TOPO II IC₅₀ for pyronaridine (32.7 μM), whereas TOPO II activity is reportedly unaffected below ~5–8 μM, consistent with a concentration‐dependent threshold for TOPO II‐mediated toxicity (Croft et al. 2012). Additionally, differences in protein sequence (≈70% identity between zebrafish and mammals) may alter drug–target interactions and help explain why developmental effects were seen in zebrafish but not in the rat study.

The developed translational mathematical exposure‐toxicodynamic model quantified the joint exposure‐toxicity relationship of cabamiquine and pyronaridine in zebrafish larvae and translated these preclinical results into the clinical perspective. The performed clinical trial simulations indicated no toxicity at the current therapeutic doses (360 mg pyronaridine and 330 mg of cabamiquine) or at the maximum clinical doses (720 mg pyronaridine and 660 of cabamiquine). Even in presence of drug–drug interactions or other scenarios that may increase the pyronaridine exposure, reaching toxic exposure is unlikely due to the high safety margin.

Few limitations regarding the modeling assumptions have to be acknowledged: The PK‐toxicity model was developed with a single species (zebrafish larvae). Due to lack of experimental data it was assumed that the fraction unbound in humans and zebrafish is similar for both drugs. Lastly, due to the lack of PK data in women, PK models from non‐pregnant women were used.

In conclusion, this comprehensive mathematical modeling of toxic interactions complements the fixed‐dose combination testing in rats, providing critical safety information across a range of exposure scenarios and dose ratios relevant to clinical development. This framework supports informed decision‐making regarding the safety of the cabamiquine‐pyronaridine combination for use during pregnancy in malaria‐endemic regions.

The favorable nonclinical safety profile, together with the translational integration of complementary preclinical approaches, supports the continued clinical development of cabamiquine‐pyronaridine combination. Furthermore, the pharmacokinetic characteristics of both compounds support the development of a single oral, once‐monthly regimen. Collectively, these findings provide compelling evidence that both cabamiquine and pyronaridine have wide developmental safety margins. The comprehensive nonclinical DART data package generated fulfills the key criteria for the ‘inflection point’ described by Stanislaus et al. (2026), supporting the proactive and ethically responsible inclusion of pregnant women in upcoming Phase 2b/3 clinical trials (Stanislaus et al. 2026). Specifically, the established NOAEL provides a robust basis for dose selection and safety monitoring in pregnant participants, while the absence of adverse drug–drug interactions at combination doses further supports the feasibility of the cabamiquine–pyronaridine regimen in this vulnerable population. Togethre, these data provide a strong scinetific foundation for advancing clinical investigation aimed at addressing the presistent malaria prevention gap during pregnancy.

Author Contributions

W.K.W., G.‐S.C. and C.D.‐G. supervised and analyzed the in vivo study. C.D.‐G. initiated the ZEDTA study and model. A.G., H.K. and T.H.H.B. conducted, acquired, and analyzed the zebrafish embryos data. A.G., S.G.W. and C.D.‐G. interpreted the ZEDTA data. S.G.W. performed the modeling of the zebrafish data and developed the clinical trial simulations. F.V. performed P.K. modeling and simulations and data interpretation. C.O. and W.K.W. contributed to the acquisition and securing of funding. C.D.‐G. wrote and edited the original manuscript. All authors contributed to the revision and editing of the manuscript.

Funding

Partial funding from the healthcare business of Merck KGaA, Darmstadt, Germany (CrossRef. funder ID: 10.13039/100009945) and from The Right Foundation (grant number RF‐2022‐T01) is gratefully acknowledged.

Conflicts of Interest

W.K.K. and G.‐S.C. are employees of Shin Poong Pharm. Co. Ltd., Ansan, Republic of Korea. C.O., T.S. and C.D.‐G. are employees of Ares Trading SA, Eysins, Switzerland, an affiliate of Merck KGaA, Darmstadt, Germany. A.G., T.H.H.B., H.K. and F.V. are employees of Merck KGaA, the healthcare business of Merck KGaA, Darmstadt, Germany. S.G.W. consultancy was funded by the healthcare business of Merck KGaA, Darmstadt, Germany.

Supporting information

Figure S1: Occurrence of “no abnormalities” in percent (y‐axis) versus pyronaridine (x‐axis) and cabamiquine exposure (nr. In facet, μM*h). The dots represent raw data, lines represent the fitted sigmoidal maximum effect model estimated only for pyronaridine considering no interaction (blue) and for the mono and combination data considering that cabamiquine potentiates the effect of pyronaridine at high concentrations while being inactive itself (red).

Figure S2: Fit plots from the pregnant rat PK model of pyronaridine. The mean concentration–time fits demonstrate that the model adequately describes the observed TK data. The predicted profiles follow the observed concentrations across the full‐time course without systematic deviations. Blue line = mean concentration, blue area = IIV (30%) on clearance (CL), bioavailability (F), and intercompartmental clearance (Q), blue point = observed data.

Figure S3: Goodness of fit plots from the pyronaridine pregnant rat PK model. The GOF plots indicate that the model adequately describes the observed data. The log–log simulation‑versus‑observation plots with the identity line demonstrate the absence of systematic bias. The residual plots show no time‑ or predictor‑dependent trends. Simulation = mean concentration, WRES = weighted residuals, red line = identity line, black points = observed data.

Figure S4: Fit plots from the pregnant rat PK model of cabamiquine. The mean concentration–time fits demonstrate that the model adequately describes the observed TK data. The predicted profiles follow the observed concentrations across the full‐time course without systematic deviations. Orange line = mean concentration, Orange area = IIV (30%) on clearance (CL), bioavailability (F), and intercompartmental clearance (Q), Orange point = observed data.

Figure S5: Goodness of fit plots from the cabamiquine pregnant rat PK model. The GOF plots indicate that the model adequately describes the observed data. The log–log simulation‑versus‑observation plots with the identity line demonstrate the absence of systematic bias. The residual plots show no time‑ or predictor‑dependent trends. Simulation = mean concentration, WRES = weighted residuals, red line = identity line, black points = observed data.

Table S1: Zebrafish embryo developmental toxicity assay endpoints analyzed at 120 hpf.

Table S2: Toxicokinetic data for cabamiquine and pyronaridine in rat embryo‐foetal development study based on non compartmental analysis (NCA).

Table S3: Summary of ovarian and uterine examination and litter observations.

Table S4: Fetal external, visceral and skeletal malformations: Fetal visceral and skeletal malformations from the rat EFD study.

Table S5: Impact of cabamiquine‐pyronaridine combination on embryo development: corresponding embryonic uptake and nominal water concentrations.

Table S6: Impact of pyronaridine on embryo development: corresponding embryonic uptake and nominal water concentrations. Mean (n = 20) substance amount per embryo after exposition with pyronaridine. Embryos were analyzed after 120 h of exposure.

Table S7: Kinetic analysis of pyronaridine concentrations in embryos from 2 to 120 hpf. Mean (n = 3) substance amount per embryo after exposition with pyronaridine at a concentration of 1 and 1.5 μM. Embryos were analyzed after 2, 8, 244, 48, 72, 96 and 120 h of exposure.

Table S8: Parameter estimates of the pregnant rat PK model of pyronaridine.

Table S9: Parameter estimates of the pregnant rat PK model of cabamiquine.

Table S10: Pyronaridine blood concentrations in fetuses and dams determined 3 h post dose on GD15 in the rat EFD study.

Data Availability Statement

The datasets analyzed during the current study are available from the corresponding author upon reasonable request and with permission of the healthcare business of Merck KGaA, Darmstadt, Germany.