Integration and application of new approach methodologies in assessing the developmental hazards: Case study with an antimalarial drug

Malaria in pregnancy remains a major public health issue capable of causing significant adverse perinatal outcomes. Due to Plasmodium sequestration in the placenta (placental malaria), and the immunosuppression during pregnancy, malaria represents a substantial risk for both the mother and fetus. Placental malaria occurs in 16%–63% of pregnant women (Darmstadt et al., 2011) with malaria and is the leading global cause of maternal anemia, low birth weight, preterm delivery, stillbirths, and miscarriages (Goldenberg et al., 2010). Each year, approximately 30–35 million women become pregnant in malaria-endemic areas of Africa, and similar numbers of women can be exposed to malaria in Asia, Oceania, and South America (Chico & Cano, 2019; World Health Organization (WHO), 2021). Unfortunately, due to a shortage of safety data on pregnant women for most of the currently recommended antimalarial drugs, there are limited treatment options for this vulnerable population, particularly in the first trimester. This highlights the need for continued improvement and development of novel antimalarial medicines with no developmental effects (teratogenicity). Over the past decades, novel antimalarials have been developed. Unfortunately, a number of these have shown to be teratogenic in preclinical animal models. Therefore, the key focus for Medicines for Malaria Venture (MMV) and partners is to incorporate embryo-fetal developmental (EFD) toxicity testing early into drug development, allowing for earlier prioritization of candidate compounds, and increasing the importance of having safe medicines for pregnancy by front-loading developmental toxicity evaluations (El Gaaloul et al., 2022). This de-risking process could act as the first step to include pregnant women and women of childbearing potential earlier in clinical trials. This population is currently excluded even though they represent a target population. In our recent study, we assessed the EFD toxicity of MMV390048, an antimalarial agent, in both rats and rabbits (Demarta-Gatsi et al., 2022). We observed adverse effects of MMV390048 on the rat fetus, specifically diaphragmatic hernia and ventricular septum defect, whereas no developmental effects were observed in rabbits. Based on these observations, MMV390048 was classified as teratogenic. In the current study, to verify the teratogenicity potential of MMV390048 and address the interspecies difference outcomes, in accordance with the 3R's (replace, reduce, refine) principles, we compared traditional EFD studies in mammals with the well-characterized new approach methodologies (NAMs; Green et al., 2016). In this study, we compared the outcomes obtained in mammalian EFDs with those observed in zebrafish (Cassar et al., 2019) and ReproTracker® (Jamalpoor et al., 2022) for the drug MMV390048. Zebrafish model has become a popular model for predicting embryotoxicity/malformations and has shown to be a good model for generating relevant information on mammalian developmental risks (Truong et al., 2014). Briefly, drug effects on zebrafish embryo development have been assessed after exposure to different concentrations (up to 100 μM) of MMV390048. However, the maximum concentration tested was limited by the compound's solubility in water. Consequently, the highest concentration achieved in fish (fish-uptake) was 8 μM at the highest concentration tested (100 μM). The second in vitro model used, ReproTracker, is a human-induced pluripotent stem cell (hiPSCs)-based biomarker assay that follows the cells differentiation during early embryonic development. Initially, cytotoxicity and solubility of MMV390048 were tested in nondifferentiated hiPSCs at different concentrations up to 1 mM. In this case, concentrations of MMV390048 were limited by cytotoxicity or solubility, and therefore the maximum noncytotoxic and/or soluble concentration (15.6 μM) was used. Next, hiPSCs were directed to differentiate toward three germ layer-specific cell types, hepatocytes, cardiomyocytes, and neural rosettes. Proper stem cell differentiation was investigated by morphological profiling and assessment of time-dependent expression patterns of cell-specific biomarkers. In this system, a decrease in the expression of specific biomarker genes and morphology disruption of the differentiated cells following compound treatment indicated teratogenicity. To allow for direct comparison between the previously published preclinical safety evaluation data for MMV390048 (Demarta-Gatsi et al., 2022) and the concentrations tested in alternative assays, we calculated and compared the maternal MMV390048 free concentrations (Cmax) in preclinical animal models with the concentrations of MMV390048 used in ReproTracker and zebrafish models (Table 1). Studies have shown that in vivo Cmax values can be compared with the in vitro concentrations and can be directly correlated to the in vitro potency values (Danielsson et al., 2000; Smith et al., 2010). In the main EFD study, MMV390048 induced diaphragmatic hernia only in one of the rat fetuses at a concentration of 13 μM (Table 1). However, in the investigative EFD study at concentrations equal to ≈20 μM, MMV390048 induced a notable developmental effect. In rabbits, MMV390048 (up to ≈10 μM, maximum tested concentration) showed no developmental effects. In the ReproTracker assay, MMV390048 up to 15.6 μM (maximum testable concentration) had no significant effect on the expression pattern of the biomarker genes nor on the functionality/morphology of all three cell types. Similarly, MMV390048 up to the highest testable concentration (8 μM) had no effect on larvae development in zebrafish (Table 1). The results of these studies suggest that MMV390048 may not be a teratogen at concentrations below 20 μM. In concordance, based on exposure data from EFD in vivo studies, IC50s ≥20 μM for individual kinases that are inhibited by MMV390048 were considered not to be relevant (Demarta-Gatsi et al., 2022). Moreover, PK/PD modeling-based simulations obtained from the Phase 1a human study predicted that a single 120 mg dose of MMV390048, resulting in a Cmax of no more than ≈3 μM in malaria naïve male individuals, would likely achieve cure in patients with Plasmodium falciparum malaria with 92% certainty (Table 2; McCarthy et al., 2020). Of note, Cmax achieved in malaria naïve male individuals can be significantly different (higher or lower) compared with individuals from endemic countries or pregnant women. In this study, we have combined traditional EFD studies in mammals with NAMs and revealed that the observed malformations of MMV390048 in rats could be related to the concentration differences or species-specific effect. Based on rat developmental studies, we expected MMV390048 to disrupt some developmental pathways in relation to diaphragmatic hernia and/or cardiovascular malformations (e.g., ventricular septum defect; Demarta-Gatsi et al., 2022) in other models, as well. However, the compound did not exhibit a teratogenic potential in any of the in vitro alternative models, nor in the rabbit study. Whether the rat or rabbit data are an outlier and their relevance to man remains an open question. The constant challenge of drug resistance and the difficulty of diagnosing malaria associated with pregnancy makes this population at risk. Therefore, MMV is working with its partners to establish in vitro screening methodologies to test antimalarials' potential teratogenicity and be able to front-load fetal/developmental toxicity studies early into drug development. In order to achieve this, a preliminary development dose range finding study will be conducted for each compound that results positive in NAMs. This study will include external fetal evaluation and visceral or skeletal examination depending on the target and expected toxicities. In this retrospective study, however, the use of NAMs would have been of limited value, as the compound had no teratogenic effect in any of the NAMs and still an EFD study at later stage of drug development would have been required. Nevertheless, this comparative approach has demonstrated that the use of NAMs can be helpful to better understand the outcome differences in animal testing and can be used for de-risking antimalarials early in drug development for teratogenicity hazards. This study further promoted MMV and partners to continue and develop a publicly supported value framework in which the degree of uncertainty is weighed up against the acceptance of the animal-free approaches. Moreover, this will allow to develop a broader normative framework for the responsible implementation of preclinical models in development and reproductive toxicology studies (DART) and make explicit possible normative presuppositions behind the current practice of (the lack of) using animal-free models in DART studies. The datasets analysed during the current study are available from the corresponding author on request and with permission of Medicines for Malaria Venture.