A target safety assessment of the potential toxicological risks of targeting plasmepsin IX/X for the treatment of malaria

Abstract

The aspartic proteases plasmepsin IX/X are important antimalarial drug targets due to their specificity to the malaria parasite and their vital role as mediators of disease progression. Focusing on parasite-specific targets where no human homologue exists reduces the possibility of on-target drug toxicity. However, there is a risk of toxicity driven by inadequate selectivity for plasmepsins IX/X in Plasmodium over related mammalian aspartic proteases. Of these, CatD/E may be of most toxicological relevance as CatD is a ubiquitous lysosomal enzyme present in most cell types and CatE is found in the gut and in erythrocytes, the clinically significant site of malarial infection. Based on mammalian aspartic protease physiology and adverse drug reactions (ADRs) to FDA-approved human immunodeficiency virus (HIV) aspartic protease inhibitors, we predicted several potential toxicities including β-cell and congenital abnormalities, hypotension, hypopigmentation, hyperlipidaemia, increased infection risk and respiratory, renal, gastrointestinal, dermatological, and other epithelial tissue toxicities. These ADRs to the HIV treatments are likely to be a result of host aspartic protease inhibition due a lack of specificity for the HIV protease; plasmepsins are much more closely related to human CatD than to HIV proteinase. Plasmepsin IX/X inhibition presents an opportunity to specifically target Plasmodium as an effective antimalarial treatment, providing adequate selectivity can be obtained. Potential plasmepsin IX/X inhibitors should be assayed for inhibitory activity against the main human aspartic proteases and particularly CatD/E. An investigative rodent study conducted early in drug discovery would serve as an initial risk assessment of the potential hazards identified.

Graphical Abstract

Introduction

Despite the increase in drug approvals in recent years, drug discovery and development remain challenging with high rates of failure [1]. In the hunt for ways to make drug discovery and development more successful, there have been several excellent analyses of why drugs fail [2, 3]. Although the reasons for failure are usually complex and multifactorial, ‘safety/toxicity’ is cited as the predominant cause of failure [2, 3] with around 40% of these failures attributable to safety issues associated with the therapeutic target itself [2]. This is not surprising since many targets that are attractive in treating disease are also likely to have important roles in normal biology and their modulation could lead to unintended consequence. So, would not it make sense to maximize our understanding of the biology of a potential drug target in order to predict and manage issues before they arise?

For many projects, the drug target may be expressed in tissues other than the intended therapeutic target; a careful analysis of expression profiles can help predict likely consequences of this and offer an insight into the benefit–risk profile. For most anti-infectives (antibiotics, antifungals, antibacterials, antiprotozoals), the target may be specific to the infectious agent suggesting that the potential for on-target toxicity is limited. However, this is not always the case since there may be related mammalian targets that need to be considered to develop an overall risk mitigation strategy. Here we explore the benefit–risk profile for a potential new target in treating malaria and propose next steps to evaluate risk.

Malaria: the problem

Malaria remains an overwhelming problem, particularly in Africa. In 2018, there were an estimated 228 million cases of malaria with an estimated 405 000 deaths from the disease globally. The burden of disease is heaviest in the World Health Organization (WHO) African Region where 93% of all malaria deaths occurred. Plasmodium falciparum (P. falciparum) is the most prevalent malaria parasite (99.7%) in the WHO African Region carried by the Anopheles mosquito, whereas P. vivax is the predominant parasite (75%) in the WHO Region of the Americas. Children aged under 5 years old and pregnant women are in the most vulnerable groups affected by malaria; children aged under 5 account for >60% of malaria deaths in 2018 worldwide [4]. Classical uncomplicated malaria presents with a combination of symptoms including fever, chills, sweats, headaches, body aches, nausea, and vomiting. Severe malarial infections are characterized by serious organ failures or abnormalities in the patient’s blood or metabolism [5]. The current fight against the disease is being waged on a variety of fronts, including the distribution of bed nets, the promotion of indoor spraying, and the development of new medicines, vaccines, and insecticides. Emerging parasite resistance to the currently available drugs is of increasing major concern.

The first instance of antimalarial drug resistance was with chloroquine. Chloroquine resistance in P. falciparum was developed independently in multiple areas of Southeast Asia, Oceania, and South America in the 1950/60s. Subsequently chloroquine resistance spread to nearly all areas of the world where P. falciparum malaria is transmitted [6]. Since then P. falciparum also developed resistance to nearly all of the other currently available antimalarial drugs, including sulfadoxine/pyrimethamine, mefloquine, halofantrine, and quinine [7]. Although resistance to these drugs tends to be much less widespread geographically, in some areas of the world, the impact of multidrug resistant malaria can be extensive [8]. Most recently, resistance to the artemisinin and non-artemisinin components of artemisinin-based combination therapy has emerged in parts of Southeast Asia, impacting the efficacy of this vital antimalarial class [9]. Chloroquine-resistant P. vivax malaria has also been identified in a number of regions including Papua New Guinea, Southeast Asia, Ethiopia, and Madagascar [6]. Therefore, there is still a continuous need for novel, differentiated approaches to treat malaria.

Plasmepsins: drug targets for malaria

As malaria routinely develops resistance to drugs, there is a continual need for novel antimalarials to combat the disease. This is especially vital in the case of multidrug resistant P. falciparum as the number of available therapies are reducing and becoming less effective. Development of small molecule plasmepsin (PM) IX/X inhibitors presents a unique opportunity to specifically target the malarial aspartic protease to prevent parasitic infection. Aspartic proteases are key contributors to the pathogenicity of P. falciparum. PMs IX and X are unique to Plasmodium and have indispensable functions in the parasite making them ideal drug targets [10–13]. Focusing on parasite-specific targets for which no human host homologue exists reduces the chance of drug toxicity. However, adequate selectivity for PMs IX/X over related mammalian aspartic proteases is critical. If there is insufficient selectivity for the Plasmodium aspartic proteases then human toxicity is to be expected, the nature of which will depend on the human aspartic proteases that are inhibited by the molecule.

In the erythrocyte stage of infection, intraerythrocytic malaria parasites degrade haemoglobin (Hb) to provide nutrients for their own growth and maturation. Hb degradation into its constituent amino acids is a multistep process involving several degradative enzymes which occurs in the acidic parasitic food vacuole (FV). The FV contains aspartic proteases, cysteine proteases, and metalloproteases, which are all believed to play roles in an ordered Hb degradation pathway [14]. Because the degradation of Hb is vital for parasite survival, strategies to inhibit this degradation pathway offers a valid approach by which to develop novel chemotherapeutic agents [11]. In the first step of degradation, Hb is broken down into large protein fragments by PM enzymes [14].

A total of 10 PMs (PMI, II, IV–X, and histo-aspartic protease (HAP)) have been identified in the genome of P. falciparum to date. PMs I–IV are transported to the parasitic FV of infected erythrocytes, where they orchestrate the degradation of Hb in a sequential manner with each PM undertaking a distinctive step in the degradation process, although they are not essential to parasite survival [15–17]. PMs expressed outside of the FV, PMV–X are expressed in all Plasmodium species [17]. PMV, PMIX, and PMX are the only other PMs expressed in the erythrocyte stages. PMV is an essential enzyme for survival as it is accountable for the cleavage of proteins that are to be exported to the host cell [18], and is therefore already a focus of drug development. PMIX and PMX have recently been highlighted as mediators of egress and invasion of the parasite further exposing the parasite at its most vulnerable state accentuating these enzymes as potential drug targets [15, 19, 20].

Within the erythrocyte the Plasmodium parasite matures from a ring to a trophozoite and then a schizont which produces merozoites primed for invasion (Fig. 1). The release of merozoites, known as egress, is a two-step process: the degradation of the parasitophorous vacuole (PV), which encloses the merozoites, and erythrocyte membrane. Invasion of a new erythrocyte takes only 10–30 s and involves recognition of the erythrocyte membrane, attachment, reorientation, and entry [21]. The proteins involved in invasion and egress are packaged into several secretory organelles in the merozoite including the rhoptries and micronemes (invasion) and the dense granule-like exonemes (egress). The activity of several serine and cysteine proteases promotes the destabilization of the PV and erythrocyte membranes which surround the parasite [22]. PMX processes subtilisin-like protease 1 (SUB1) to activate it. Inhibition of PMX results in the accumulation of SUB1 precursor [10]. Mature SUB1 is required for the degradation of both the PV and erythrocyte membrane to allow the dissemination of merozoites from a mother schizont. To initiate the egress cascade, SUB1 activates cysteine proteases called SERAs and merozoite surface proteins called MSPs [23]. Full block of PMX traps parasites within the PV membrane, whereas partial block allows egress from this membrane but prevents escape from the erythrocyte membrane. This suggests a higher level of activated SUB1 is required for its effects on the erythrocyte [10]. Invasion also crucially relies on serine proteases to activate or remove ligands involved in interactions with the host erythrocyte [24].

Figure 1.

Malaria infection cycle. The Anopheles mosquito bites a human taking a blood meal and sporozoites from the salivary gland enter the human blood stream. The sporozoites move to the liver and invade hepatocytes. Within the liver they develop to produce exoerythrocytic merozoites which are released into the blood stream. Merozoites invade erythrocytes which takes between 10 and 30 s and involves recognition of the erythrocyte membrane, attachment, reorientation, and entry. Within the erythrocyte merozoites grow into a ring or trophozoite and mature in to a schizont. The schizont produces merozoites which are primed to invade further erythrocytes. The release of merozoites, known as egress, is a two-step process: the degradation of the parasitophorous vacuole, which encloses the merozoites, and the degradation of the erythrocyte membrane. Gametocytes are formed from the asexual blood stage and are taken up by a feeding mosquito. The gametocytes mature in the mosquito gut to become male and female gametes. The fertilized zygote develops to an ookinete and an oocyst inside the mosquito. Sporozoites then migrate to the salivary glands where they can be injected into a new human host to propagate the spread of infection. Adapted from Cowman et al. [21] and Weiss et al. [24].

PMIX and PMX are expressed in mature blood-stage schizonts and invasive merozoites and fulfil indispensable functions. PMIX and PMX are specifically involved in these egress-invasion processes evidenced in two recent studies by Pino et al. [25] and Nasamu et al. [11]. Both studies used similar approaches to characterize PMIX and PMX, supporting their validation as drug targets whilst also presenting two novel inhibitors 49c and CWHM-117. Both groups were unable to generate parasite knockouts, inferring essential survival functions associated with PMIX/X. The groups reported that the compounds lead to egress and invasion phenotypes similar to those seen in the knockdown parasites. When 49c was removed before parasite egress, the parasite invaded new erythrocytes normally indicating that treatment, at least with this drug, is time dependent. For 49c to inhibit egress, it needs to be present for an extended period, to enable it to block the gradual accumulation of mature SUB1 to an extent sufficient to reduce SUB1 activity below the threshold required for membrane rupture. Crucially, this impacts on the required pharmacokinetic (PK) properties of PMIX/X inhibitors, which may require long plasma half-lives in vivo and/or long residence times in order to exert the required physiologically significant impact. In silico structural and physicochemical inspection of 49c highlighted poor PK properties emphasizing the need for the development of tailored inhibitors with desirable therapeutic properties against Plasmepsin IX/X [26]. 49c was also tested in vivo on other Plasmodium stages with some efficiency against hepatocyte and gametocyte egress as well as interference with the processing of the cell traversal protein for ookinetes and sporozoites in the ookinete. Therefore, the absence/inhibition of PMIX and PMX blocks parasite development by impairing egress or invasion of a new erythrocyte. Oral treatment of a novel dual PMIX/X inhibitor WM382 cured mice of the rodent Plasmodium parasite P. berghei and prevented blood infection from the liver. In addition, WM382 was efficacious against P. falciparum asexual infection in humanized mice and prevented transmission to mosquitoes [10]. No toxicological endpoints were evaluated in any of these studies.

Potential toxicities of inhibiting mammalian aspartic proteases

Target-related safety issues are responsible for many drug project failures [2], therefore, evaluating the potential for on target-related toxicity, in conjunction with developing understanding of efficacy, can aide in decision making through the drug discovery process. In this case, the proposed targets PMIX/X are specific to Plasmodium species; nonetheless there is the potential for drugs that target PMIX/X to also inhibit human aspartic proteases due to their homology. Mammalian aspartic proteases include the digestive enzymes (pepsin), the intracellular cathepsin D (CatD) and cathepsin E (CatE), and renin. Aspartic proteases are important in hydrolytic processes and have become drug targets for several diseases including malaria. For instance the inhibition of renin in hypertension treatment [27], human immunodeficiency virus (HIV) aspartyl protease inhibitors for autoimmune deficiency syndrome (AIDS) [28], CatD targeting antibodies in breast cancer [29], and BACE1 inhibitors for Alzheimer’s disease [30]. Initially, aspartic proteases are synthesized as proenzymes (zymogens) and are activated upon cleavage of the pro-segment [31]. Aspartic proteases belong to a class of protease enzymes which use a catalytic aspartic acid dyad to cleave peptide tetrahedral intermediate. The catalytic aspartic acid residues take on different roles in a general acid–base or ‘push and pull’ reaction [32]. In spite of the heterogeneity among aspartic proteases, the active site (motif Asp-Ser/Thr-Gly) and mechanism of action is conserved throughout the family [32–34]. The high levels of conservation between Plasmodium and mammalian aspartic proteases has directed lead optimization studies to focus on improving PM inhibition selectivity versus the related human aspartic proteases [35].

The potential toxicity risks associated with mammalian aspartic protease inhibition were elucidated through conducting a thorough assessment of the literature and publicly available data, to produce a target safety assessment (TSA) [33, 36, 37]. Though primarily an in cerebro exercise, the goal of a TSA is to identify potential unintended consequences of target modulation and to propose a risk evaluation and mitigation strategy to assist in early program advancement and to anticipate, monitor, and manage potential clinical adverse events (AEs) [33, 36, 37]. TSAs bring together target homology information, gene and protein expression profiles, transgenic and mutation phenotype data, including loss- or gain-of-function within mouse/rat and human phenotypes, and understanding of the role of the target under ‘normal’ physiological conditions, as well as in disease state. By combining these data with any public domain information on competitor compounds, key risks can be identified and categorized by organ/tissue and/or physiological functions as the basis of a risk mitigation plan with ranking of risks and potential next steps [38]. For this specific TSA, data were collated on in vitro investigations and in vivo animal studies highlighting mammalian aspartic protease function together with actual adverse drug reactions (ADRs) noted in patients treated with approved aspartic protease inhibitors (Fig. 2).

Figure 2.

Potential risks of mammalian aspartic protease inhibition. An in cerebro and in silico target safety assessment of publicly available literature was performed to uncover the possible toxicity risks to organ systems in the body. As plasmepsins are aspartic proteases specific to Plasmodium, the associated toxicity risks of targeting them would arise from nonspecific host aspartic protease inhibition. A compound with a lack of specificity to plasmepsin IX/X may interact with mammalian aspartic proteases due to their structural similarity, potentially leading to toxicity. Based on the known adverse reactions to existing compounds, along with in vitro and in vivo data, the probability of organ toxicity occurrence was determined. A scale of low (green box), medium (amber boxes), or high (red boxes) was used to categorize the probability of toxicity occurrence. The impact of possible toxicities to progression of the compound if they did indeed occur was assessed based on the type and severity of the effect on target organ inhibition.

Cathepsin D

CatD is a mammalian aspartic endo-protease that is ubiquitously distributed in lysosomes [39]. CatD degrades proteins and activates precursors of bioactive proteins in pre-lysosomal compartments [40]. The many physiological functions of CatD include metabolic degradation of intracellular proteins, activation and degradation of polypeptide hormones and growth factors, activation of enzymatic precursors, processing of enzyme activators and inhibitors, brain antigen processing, and regulation of apoptosis [41–44]. CatD can also be found in the extracellular space [45]. CTSD is the gene that encodes CatD; homozygous deletion of Ctsd in mice leads to progressive atrophy of the intestinal mucosa, increased apoptosis in the thymus and profound destruction of lymphoid cells with early lethality in the postnatal phase [46]. These data indicate that CatD is required in certain epithelial cells for tissue remodelling and renewal. Human deficiency of CatD has been reported as an underlying cause of congenital human neuronal ceroid-lipofuscinosis (NCL). NCL is a lysosomal storage disease that results from excessive accumulation of lipopigments (lipofuscin) in the body’s tissues. It is a devastating neurodegenerative disorder characterized by severe neurodegeneration, developmental regression, visual loss, epilepsy, and premature death [47, 48]. CatD inhibitors have yet to reach the clinic; mice used for preclinical xenograft models of CatD targeting antibodies showed no overt toxicity as indicated by an absence of weight loss although no specific toxicity endpoints were measured [25].

Several structurally distinct beta-secretase (BACE1) inhibitors have been withdrawn from development after inducing ocular toxicity following chronic treatment in animal models; CatD was identified as a principal off-target of these BACE1 inhibitors in human cells. Quantitation of CatD target engagement in cells has been shown to be predictive of ocular toxicity in vivo, suggesting that off-target inhibition of CatD is a principal driver of ocular toxicity for BACE1 inhibitors [49]. Therefore, inhibition of CatD could potentially result in a range of toxicological effects including epithelial cell remodelling and renewal which result in toxicity to the GI tract, the eye, and congenital abnormalities. As CatD has such a broad physiological role in many tissues it is difficult to predict which target organ of toxicity would predominate with CatD inhibition.

Cathepsin E

CatE is another major mammalian intracellular aspartic proteinase implicated in the physiological and pathological degradation of intracellular and extracellular proteins. It is an intracellular non-lysosomal glycoprotein that is mainly found in the skin and in immune cells. It is found at highest abundance on stomach epithelial mucus-producing cell surfaces [50]. CatE also plays an important role in immune responses since it is implicated in antigen processing via the major histocompatibility complex (MHC) class II pathway [50]. Mice with homozygous knockout of Ctse, the CatE gene, show an increased susceptibility to bacterial infection associated with decreased expression of multiple cell surface Toll-like receptors [51]. Deficiency of CatE in mice has also been shown to cause atopic dermatitis (AD), a pruritic inflammatory skin disease [52]. The reduced expression of CatE was also observed in erythrocytes of both humans with AD and in an AD mouse model indicating that CatE deficiency might therefore be linked to the induction of AD. CatE deficiency in homozygous knockout mice also induces a form of lysosomal storage disorder characterized by accumulation of lysosomal membrane sialoglycoproteins (glycoproteins which contain sialic acid as one of their carbohydrates), and the elevation of lysosomal pH in macrophages. These striking features were also found in wild-type macrophages treated with pepstatin A (a potent inhibitor of aspartyl proteases), and Ascaris inhibitor (a strong inhibitor of CatE but not CatD) [53]. These results suggest that CatE is important for preventing the accumulation of sialoglycoproteins that can induce a form of lysosomal storage disorder. Therefore, as CatE plays an important role in the maintenance of immune system homeostasis by participating in host defence mechanisms, inhibition of CatE could potentially result in increased incidence of infection and dermatitis. Furthermore, CatE inhibition could potentially lead to a lysosomal storage disorder in macrophages. However, as the aim of this project is to develop a PMIX/X inhibitor that functions as a single dose treatment, the risk of these toxicities is significantly reduced.

Renin

Renin (angiotensinogenase) is a mammalian aspartic protease protein and enzyme secreted by the kidneys. It participates in the body’s renin–angiotensin–aldosterone system (RAAS) that mediates the volume of extracellular fluid, and arterial vasoconstriction, thereby regulating the mean arterial blood pressure [31, 54]. Renin is secreted from specialized granular cells found in the juxtaglomerular apparatus of the kidneys, and is secreted in response to decreased arterial blood pressure, decreased sodium chloride levels, and sympathetic nervous system activation. Renin is classically termed as a hormone, although renin is an enzyme responsible for the hydrolysis of the amide bond at L10–V11 of angiotensinogen (from the liver) to angiotensin I which is converted to the potent vasoconstrictor angiotensin II [55, 56]. Angiotensin II leads to the constriction of blood vessels, increased secretion of antidiuretic hormone and aldosterone; all causes an increase in blood pressure. Renin’s primary function is therefore to cause an increase in blood pressure, leading to restoration of perfusion pressure in the kidneys. There are several renin inhibitors and angiotensin II receptor antagonists on the market for the treatment of a range of cardiovascular disorders, including hypertension and heart failure, nephropathy (diabetic and nondiabetic), and atherosclerosis. They are safe and well tolerated in these patient populations [27]. Inhibition of renin could potentially result in a diminished response to decreased arterial blood pressure thereby leading to hypotension, and reduced perfusion pressure in the kidneys which could potentially result in acute kidney injury.

Pepsin

Pepsin is the major enzyme present in the mammalian stomach and is the first protease that food proteins encounter in the digestive tract. It is involved in protein digestion and its proposed physiological function is the stimulation of the disintegration of the food bolus by fast hydrolysis of intact proteins, rather than finely digesting the proteins into absorption-ready peptides [57]. There is no evidence in the literature from transgenic animal models or human mutation phenotypes to highlight the effect of pepsin inhibition on normal biology. However, inhibition of pepsin could potentially result in a decrease in the stimulation of digestion which could lead to a decrease in gastric emptying and nutrient absorption.

Napsin A

Napsin A is an aspartic protease present in the epithelial cells of the mammalian lung and kidney [58]. Expression is seen in type II alveolar cells of the lung [59], where it is involved in the processing of surfactant protein B (SP-B) [60], and in the renal proximal tubules where it functions as a lysosomal protease, highlighting a role in protein catabolism [61]. SP-B plays a critical role in the functioning of healthy lungs; its absence leads to lung conditions, the most common of which is acute respiratory distress syndrome [62]. Inhibition of Napsin A could potentially lead to protein overload in proximal tubules, which is ultimately known to lead to tubulointerstitial damage [63]. Therefore, inhibition of Napsin A could potentially result in lung or kidney toxicity.

Beta-secretase enzyme

There are two forms of the beta-secretase enzyme, BACE1 and BACE2, which were initially identified as transmembrane aspartyl proteases cleaving the amyloid precursor protein (APP) at the β-site [64]. BACE1 is highly expressed throughout the brain, whereas BACE2 is expressed at low levels in the brain but expressed at varying levels in peripheral tissue [31]. BACE1 is an aspartic-acid protease important in the formation of myelin sheaths in peripheral nerve cells [65]. Due to the cleavage specificity and localization of BACE1 activity, BACE1 is believed to be the secretase responsible for the formation of plaques in the brain. BACE cleaves the APP at the β-secretase site, a critical step in the Alzheimer’s disease pathogenesis. Comparison of BACE1 to other aspartic proteases such as CatD and CatE, Napsin A, pepsin, and renin revealed little similarity with respect to the substrate preference and inhibitor profile. On the other hand, these parameters are all very similar for the homologous enzyme BACE2 [66].

BACE1- and BACE2-deficient mice demonstrate a wide range of physiological substrates and functions for both proteases both within and outside of the nervous system. For BACE1 this includes axon guidance, neurogenesis, muscle spindle formation, and neuronal network functions, whereas BACE2 has been shown to be involved in pigmentation and pancreatic β-cell function [64]. It is likely that many of the developmental BACE1 functions, such as myelination, would not be affected in adult patients, with the impact of inhibition occurring in young children. However, neurogenesis and axon guidance in the hippocampus would be relevant for all patient populations. Muscle spindle maintenance is an adult BACE1 function and is compromised in adult mice treated with a BACE inhibitor [67]. Therefore, inhibition of BACE1 could compromise brain function and muscle spindle maintenance whereas inhibition of BACE2 could potentially result in hypopigmentation or a perturbed pancreatic β-cell function.

Toxicities associated with approved aspartic protease inhibitors

One of the most well-known protease-based therapies is the HIV aspartic protease inhibitors, which are widely used in the combined treatment of AIDS. These inhibitors block the crucial viral maturation stage and thereby reduce the spread of HIV. Such inhibitors approved by the FDA include saquinavir, ritonavir, indinavir, nelfinavir, lopinavir, atazanavir, amprenavir, fosamprenavir, tipranavir, and darunavir [28]. It has been shown that saquinavir, ritonavir, and indinavir, at therapeutically relevant concentrations, inhibit the growth of P. falciparum in vitro cell cultures [68, 69] and in vivo in mice [70, 71].

AEs in response to treatment with the main subclasses of highly potent antiretroviral therapy have recently been reviewed [72]. Aspartic protease inhibitors resulted in the largest number of reported AEs caused by continuous use. Most AEs were related to hyperlipidaemia but other AEs included renal, gastrointestinal, and dermatological toxicities [73, 74]. In one case, aspartic protease inhibitors were linked to the development of anaemia in post-partum women [72].

The main AEs of ritonavir include vomiting, nausea, and abdominal pain at onset of treatment. An increase in the concentration of ritonatir has also been associated with hypertriglyceridaemia [72]. AEs reported after 48 weeks of treatment with tipranavir/ritonavir include gastrointestinal symptoms and grade 3–4 elevations in cholesterol levels [75]. The use of atazanavir/ritonavir also resulted in a less favourable lipid profile with a significant increase in plasma triglyceride levels [76].

Clinical trials with HIV aspartic protease inhibitors show that major AEs can occur during treatment, necessitating interruption and/or changes of medication. The AEs are thought to be due to the inhibition of human aspartic proteases such as CatD. For example, interactions between ritonavir and nonretroviral aspartic proteinases are well established, with Ki values against human aspartic proteases CatD and CatE of 20 and 8 nmol/l, respectively. The plasmepsins are much more closely related to human CatD than to HIV proteinase so if development of antimalarial therapy based on HIV proteinase inhibitors is considered, the potential interactions with the human aspartic proteinases may need to be minimized [77].

Nonclinical safety assessment plan

A nonclinical safety assessment plan was formulated to test the selectivity versus the intended therapeutic target of compounds progressing in the programme (Fig. 3). The chemistry of the compound should be driven by prioritizing selectivity against CatD and CatE inhibition since these are likely to have the most undesirable and severe effects.

Figure 3.

Non-clinical safety assessment plan. A non-clinical safety assessment plan was formulated to test the selectivity of compounds progressing in the programme for plasmepsin IX/X versus mammalian aspartic proteases. First in the selectivity testing cascade, a high-throughput assay to understand direct inhibition of mammalian aspartic proteases in comparison to plasmepsins should be performed. Then a cell-based assay would provide a more accurate reflection of compound potency. Combining in vitro IC₅₀ values with pharmacokinetic data to model target engagement would further develop understanding of the specificity profile. A comparison of lead compounds with compounds known to interact with mammalian aspartic proteases may also be useful in guiding the chemistry. An investigative in vivo toxicity study, in combination with an understanding of exposures required for efficacy, would be useful to identify and understand the selectivity window against mammalian targets required. Endpoints for the in vivo toxicity study would be chosen based on the findings from the TSA and in vitro studies. Before the first-in-human trial, an assessment in both a rodent and non-rodent species for a duration sufficient to support the initial clinical protocol will need to be conducted. A combination of a range of in vitro and in vivo studies is crucial to understand the overall safety profile and ensure progression of the compound(s) in the programme.

Initially, an isolated enzyme assay could be used as a high-throughput screen to assess direct inhibition of mammalian aspartyl proteases, including CatD and CatE. Translating these inhibition values to the context of plasmepsin potency can be challenging, due to technical differences between assay set-up such as platforms used, concentrations of protein and substrates. Thus, a direct comparison may not be a true reflection of the selectivity profile. However, when designing the compounds, understanding the selectivity window in these assays sufficiently to drive it wider with each chemistry optimization round is a pragmatic approach. It is not possible to predict adverse outcomes from the isolated biochemical enzyme in vitro IC₅₀ value alone, as the duration of inhibition (both in terms of hours in the day and number of days) will determine the incidence, severity, and reversibility of toxicity. From the in vivo pharmacokinetic (PK) data, it may be possible to predict levels of off-target engagement, and also get some understanding of likely toxicities, as described by Zuhl et al. [49].

A lower throughput, more quantitative method may have the potential to provide a more accurate determination of inhibitor potency and could be used to assay newly identified targets in live cells [49]. In their assay Zuhl et al., reported that several BACE1 inhibitors blocked activity of CatD in a cell-based assay to a much greater potency compared to the isolated enzyme assay, indicating that results in the more sophisticated assay are dependent on more than just the kinetics of enzyme inhibition. This can also happen in the reverse, with cellular potency dropping off compared to isolated enzyme potency, due to factors such as cellular efflux and/or permeability. Another approach is to design a reporter substrate that is primarily cleaved by the target protease in a complex cellular extract or even in intact cells [78]. Again, however, it is not possible to predict adverse outcomes from the cell based in vitro IC₅₀ value alone, as it does not take into account the ADME (absorption, distribution, metabolism, or excretion) properties of a compound, all of which can influence the toxicological profile. Combining in vitro IC₅₀ values with PK data and modelling possible target engagement and outcomes in comparison to compounds known to cause the toxicity by this mechanism may be useful.

Cell-free and in vitro assays are useful to understand selectivity and predict what may happen in vivo. But ultimately, an in vivo assessment in the rodent early in the programme, when there are tool compounds with suitable in vivo PK properties, would help identify and understand the selectivity window against mammalian targets that may be required in context with exposures required for efficacy. Both rat and mouse aspartic proteases have a high level of homology with the human orthologs on a gene and protein level (data not shown). Mammalian aspartic proteases also share topologically similar domains and active site residues [33]. Together this suggests that the rodent enzymes are sufficiently similar in catalytic site structure and function to be able to make useful selectivity comparisons for small molecule compounds. Endpoints to consider (depending on the enzyme selectivity profile if known) include cholesterol and triglyceride levels, visual acuity assessment and assessment of accumulation of autofluorescent material in the retinal pigmented epithelium (RPE), histopathology of major organs (stomach, intestine, thymus, skin, kidney, lungs, liver), CNS functional endpoints, blood pressure assessment, and potentially assessment of aspartic protease levels in tissues with lesions. Based on the predicted toxicities, special stains such as Schmarl’s stain for lipofuscin may be required. Even with a well-designed experiment it could be difficult to capture all these endpoints in a single experiment, therefore, depending on the selectivity profile of the compounds being tested, prioritization of endpoints may be required.

Depending on the overall selectivity profiles of the compounds, the results from an initial in vivo investigative assessment, the clinical plan, the planned patient population, and the overall confidence in and appetite for risk within the programme, decisions on the progression of a compound(s) for further evaluation can be made. Before a candidate drug is administered to humans, assessment in both a rodent and non-rodent species for a duration sufficient to support the initial clinical protocol will need to be conducted, alongside functional assessment of the major organ systems (nervous, respiratory, and cardiovascular). As the compound progresses through clinical phases, further preclinical studies will need to be conducted to support administration in women of child-bearing potential and children, key patient populations for an antimalarial therapy.

Discussion

Malaria remains a highly challenging problem with >200 million cases annually. Since the malaria parasite routinely develops resistance to drugs, there is a continual need for novel antimalarials to combat the disease. Development of a small molecule plasmepsin IX/X inhibitor presents a unique opportunity to specifically target the malarial aspartic proteases. Furthermore, focusing on parasite-specific targets for which no human host orthologue exists reduces the chance of drug toxicity. However, since this is a small molecule programme rather than a vaccine or antibody approach, there may be lack of specificity, introducing the potential for risks associated with hitting related mammalian targets.

We identified a range of mammalian proteins that inhibitors of plasmepsin IX/X may interact with including pepsin, cathepsins, β secretases, and renin. This reinforced the need for specificity and provided the necessary data to direct the chemistry early in the programme. By assessing gene and protein expression, transgenic mouse data, human mutation phenotypes, and ADRs, we have been able to identify the key risks associated with the project. The risk evaluation and mitigation strategy suggest that a thorough evaluation of the selectivity profile coupled with an early in vivo evaluation will be vital to anticipate, monitor, and manage potential clinical AEs. Additionally, since the project aim is to develop a single-dose treatment, such a study will be important to help define onset, dose-dependency, reversibility, and margin of safety.

Conclusion

Compared with TSAs conducted for drug projects in many other therapy areas, TSAs for malaria differ since there is often no identical mammalian target. However, the overall conclusion remains the same as for other therapy areas; use all available data to generate an early risk management plan that can identify and mitigate any potential toxicological issues in the project.

Conflict of interest statement

RR is a director and cofounder of ApconiX, an integrated toxicology and ion channel company that provides expert advice on non-clinical aspects of drug discovery and drug development to academia, industry, and not-for-profit organizations. RR, JB, CS and PS are employees of ApconiX. Clients of ApconiX include UCB Biopharma SRL and MMV. JPV is an employee and stockholder of UCB Biopharma SRL.