Tafenoquine: A Step toward Malaria Elimination

Abstract

There is a pressing need for compounds with broad-spectrum activity against malaria parasites at various life cycle stages to achieve malaria elimination. However, this goal cannot be accomplished without targeting the tenacious dormant liver-stage hypnozoite that causes multiple relapses after the first episode of illness. In the search for the magic bullet to radically cure Plasmodium vivax malaria, tafenoquine outperformed other candidate drugs and was approved by the U.S. Food and Drug Administration in 2018. Tafenoquine is an 8-aminoquinoline that inhibits multiple life stages of various Plasmodium species. Additionally, its much longer half-life allows for single-dose treatment, which will improve the compliance rate. Despite its approval and the long-time use of other 8-aminoquinolines, the mechanisms behind tafenoquine’s activity and adverse effects are still largely unknown. In this Perspective, we discuss the plausible underlying mechanisms of tafenoquine’s antiparasitic activity and highlight its role as a cellular stressor. We also discuss potential drug combinations and the development of next-generation 8-aminoquinolines to further improve the therapeutic index of tafenoquine for malaria treatment and prevention.

Graphical Abstract

Malaria continues to pose a tremendous financial and health burden to the world, accounting for 228 million cases and 405000 deaths in 2018.¹ The spreading resistance to available antimalarial drugs and the reduced susceptibility to front-line artemisinin-based combination therapies have slowed the progress of malaria control and elimination. New drugs with a broad-spectrum activity against multiple malaria life cycle stages in various species, including those with multidrug resistance, are desperately needed to overcome emerging drug resistance. These properties are also requisite for an antimalarial agent to satisfy the global target of eradicating malaria.

Among the human-infective species, Plasmodium falciparum and Plasmodium vivax are the main culprits in most malaria cases. These apicomplexan parasites are transmitted by female Anopheles mosquitoes when taking a blood meal in which sporozoites (a liver-invasive parasite form) enter the blood circulation and invade hepatocytes. Following the liver-stage schizogony (asexual reproduction by multiple nuclear divisions before cytokinesis), thousands of parasites are released into the bloodstream and invade red blood cells (RBCs). At this time, the merozoites (an RBC-invasive parasite form) develop into blood-stage schizonts that produce 10−36 daughter merozoites capable of further red cell invasion and expansion. During the blood stage, a proportion of merozoites will differentiate into male and female gametocytes which, once taken up by another mosquito, fuse to form zygotes and eventually give rise to sporozoites in the salivary glands of Anopheles mosquitoes.

Unlike P. falciparum, a fraction of P. vivax sporozoites can enter a dormant liver stage that causes malaria relapse within a range of weeks to several years following the initial episode of illness. This difficult-to-treat dormant liver stage is attributed to the latent parasite form called a hypnozoite, which is refractory to most antimalarial medications. Mathematical models indicate that the hypnozoite reservoir continues to fuel malaria transmission, and therefore, the goal of elimination cannot be achieved without targeting this reservoir.²⁻⁴ In 2018, the U.S. Food and Drug Administration and the Australian Therapeutic Goods Administration approved tafenoquine for the radical cure (clearing of all parasite forms, including hypnozoites) of P. vivax infection (Krintafel, Kozenis) and for malaria chemoprophylaxis (Arakoda, Kodatef) (Figure 1).⁵ Before tafenoquine’s approval, the community had been counting on primaquine as the sole antirelapse therapy since its registration in 1952.⁶ Like primaquine, tafenoquine is an 8-aminoquinoline derivative that shows similar tolerability and efficacy in preventing recurrence of P. vivax malaria,⁷⁻⁹ while in actual clinical settings, tafenoquine is expected to have a higher efficacy due to its superior pharmacokinetic properties. This important feature allows for single-dose treatment, a significant improvement to the standard 14-day primaquine regimen that results in a low compliance rate.¹⁰⁻¹³ Additionally, tafenoquine has a broad-spectrum activity against liver- and blood-stage schizonts and gametocytes in both P. falciparum and P. vivax, making this drug suitable for malaria prophylaxis (Figure 2).^7,14,15 Recently, the evolution of 8-aminoquinolines and the emergence of tafenoquine have been described in detail.⁷ In this Perspective, we discuss the mechanisms behind tafenoquine’s activity and toxicity, especially from the point of view of drug-induced cellular stresses. We also discuss potential drug combinations and the development of next-generation 8-aminoquinolines for extending the therapeutic window of tafenoquine.

Figure 1.

Structures of tafenoquine and other 8-aminoquinolines.

Figure 2.

Broad-spectrum activity of tafenoquine on different life cycle stages of Plasmodium parasites. Plasmodium parasites are transmitted by Anopheles mosquitoes and undergo the (A) liver stage and (B) blood stage in humans. (C) The blood-stage gametocytes are then taken up by another mosquito and eventually develop into the liver-infective form to complete the life cycle. Yellow arrows indicate parasite development (hypnozoite) in P. vivax (Pv) but not in P. falciparum (Pf). TQ, tafenoquine; PQ, primaquine.

MECHANISMS OF ACTION: WHAT IS STRESSING THE PARASITES OUT?

The mechanism of action of tafenoquine remains largely unknown, and currently, there are no known molecular targets for this drug or other 8-aminoquinolines.¹⁶⁻¹⁸ Previous clinical data suggest a link between cytochrome P450 isozyme 2D6 (CYP 2D6) and the metabolism/therapeutic activity of primaquine.¹⁹⁻²¹ This connection was supported by a CYP 2D knockout mouse model in which primaquine failed to prevent Plasmodium berghei infection.²² The causal prophylactic activity (activity against the liver stage of infection) of primaquine was then restored when the human CYP 2D6 gene was introduced.²² Similar experiments suggest that the antiplasmodial activity and pharmacokinetics of tafenoquine require metabolic activation by CYP 2D enzymes.^23,24 A more extensive genetic association study could address if the enzymatic activities of CYP 2D6 and other metabolic enzymes dictate the curative efficacy of tafenoquine in humans.^14,25

During the dormant stage, Plasmodium parasites generally suppress metabolic and transcriptional activities while exhibiting transcriptionally active redox metabolism and heat shock protein 70 expression, which renders resilience against most antimalarial agents.²⁶ However, tafenoquine and other 8-aminoquinolines seem to have the capability to subvert cellular homeostasis in established hypnozoites. Through the CYP 2D6-mediated metabolic pathway, primaquine is metabolized to multiple hydroxylated species that are unstable and highly redox reactive.^7,27,28 These hydroxylated species are transformed into similarly reactive quinoneimine metabolites that can be converted back to the hydroxylated derivatives under aerobic conditions, causing redox cycling and accumulation of hydrogen peroxide (H₂O₂).^27,29 The buildup of reactive oxygen species (ROS) in parasite-infected hepatocytes may eventually lead to parasite death.^5,27,30 The same mode of action was recently found in the bone marrow where primaquine metabolism generated H₂O₂ to kill gametocytes and blood-stage schizonts.³⁰⁻³² Given that the antiplasmodial activity of tafenoquine may require CYP 2D activation, production of its reactive metabolites and H₂O₂ is likely to be involved. Thus far, the most recognized primaquine metabolite responsible for the activity is 5-hydroxyprimaquine, the presence of which is detected by measuring the downstream 5,6-o-quinone product.^27,33,34 Unlike primaquine, tafenoquine has a 3-(trifluoromethyl)phenoxy group that prevents position 5 from direct oxidation, conferring the slow elimination and long-acting activity of tafenoquine.³⁵ There was thus speculation that O-dealkylation and oxidation of the 2- and 6-methoxy groups could give rise to the reactive quinoneimine metabolites.³⁵ However, the 5,6-o-quinone metabolite of tafenoquine was also detected in vivo, suggesting the presence of 5-hydroxylated species (Figure 3).^24,36 A more complete profiling of tafenoquine metabolism and the enzymes involved would help elucidate its mechanism of action and potentially facilitate future drug design.

Figure 3.

Hypothesized mechanisms underlying the activity and toxicity of tafenoquine. Dashed arrows indicate the production of a hypothetical tafenoquine metabolite via G6P 2D-dependent or -independent pathways.

Redox cycling of hydroxylated primaquine metabolites also generates semiquinonimine radical that may target nucleophiles as well as alkylate parasite proteins and/or other biomolecules.^7,27,29,37 However, this potential mechanism remains largely unexplored. Intriguingly, endoperoxides such as artemisinin are potent antimalarial drugs that form carbon-centered radicals to alkylate parasite proteins and unsaturated membrane lipids.^38,39 Like 8-aminoquinolines, the mechanism of action of endoperoxides and their molecular targets have long been elusive. Chemoproteomic studies using artemisinin probes attached with clickable alkyne and azide tags have revealed more than one hundred P. falciparum parasite proteins modified by artemisinin.⁴⁰⁻⁴² Similar strategies could be applied to identify protein adducts, if any, of tafenoquine and primaquine. To date, no genetic marker has been found for tafenoquine or primaquine resistance.^16,17 Such “resistance” is difficult to determine due to treatment adherence (for primaquine), co-administration with blood schizonticides, and the varying CYP 2D6 activity in the population.²³ If parasites with higher tafenoquine or primaquine tolerability could be raised under drug selection, it could generate testable hypotheses about molecular targets. Unfortunately, P. vivax long-term culture has not been achieved given its restricted tropism to invade reticulocytes.^35,43,44 Perhaps the recently established Plasmodium cynomolgi continuous culture system⁴⁵ and tafenoquine’s ability to inhibit the P. falciparum asexual blood stage⁷ will facilitate future studies to help elucidate the mechanism(s) underlying tafenoquine’s activity and resistance. Because this compound triggers oxidative stress (and potentially proteotoxic stress) in Plasmodium parasites, a systemic change in the parasite stress response machinery may confer tafenoquine tolerance in the parasites.

Deletion of the CYP 2D gene cluster, as described, abolished the causal prophylactic activity of tafenoquine in mice,^23,24 yet this genetic deletion did not affect the ability of tafenoquine to inhibit blood-stage schizonts and gametocytes, suggesting a CYP 2D-independent activity for these stages.⁴⁶ However, this observation does not necessarily indicate a different mode of action for the blood stage. Fasinu and colleagues⁴⁷ recently showed that primaquine incubated with human RBCs could transform into the 5,6-o-quinone species. Therefore, a non-enzymatic activation of 8-aminoquinolines may exist. Other than ROS production, tafenoquine also caused mitochondrial dysfunction accompanied by increased intracellular Ca²⁺ levels in the related protozoan parasites Leishmania and Trypanosoma brucei.^48,49 Additionally, tafenoquine may trigger eryptosis (programmed cell death of erythrocytes) in Plasmodium-infected RBCs.^50,51 Whether these events are downstream of cellular stresses induced by tafenoquine or direct targets of this drug is currently unknown.

MECHANISMS BEHIND TAFENOQUINE’S TOXICITY

The major safety concerns with tafenoquine and other 8-aminoquinolines are their hemolytic toxicity in glucose-6-phosphate dehydrogenase (G6PD)-deficient individuals and the general hematotoxicity (primarily methemoglobinemia) in all patients.⁷ Methemoglobinemia is characterized by an increased level of methemoglobin that carries oxidized ferric iron (Fe³⁺) and thus impairs the protein’s oxygen binding capacity.⁵² Although the elevated methemoglobin level showed no clinical symptoms in tafenoquine-treated patients, severe methemoglobinemia could occur in certain clinical settings such as G6PD deficiency.^8,9,53,54 As methemoglobin formation involves hemoglobin oxidation, many oxidizing agents have been recognized as strong inducers of methemoglobinemia.⁵⁵ Not surprisingly, the oxidative activity of the hydroxylated 8-aminoquinoline metabolites is the main culprit for this adverse event.⁵⁶⁻⁵⁸ Computational analyses further revealed the role of these metabolites as electron donors for hemoglobin-bound oxygen, which then generates H₂O₂ and methemoglobin.⁵⁹⁻⁶¹

Tafenoquine can cause severe hemolytic anemia in G6PD-deficient individuals. G6PD deficiency is an inherited X-linked genetic disorder with diverse single-nucleotide polymorphisms that affect G6PD enzyme activity to a varying degree. This genetic abnormality is especially widespread in malaria endemic regions (8% on average with >30% in some malarious countries) and thus affects a relatively large proportion of the population susceptible to malaria infection.⁶²⁻⁶⁴ As with primaquine, the hemolytic potential of tafenoquine appears to correlate with G6PD enzyme activity in humans.^65,66 Thus, G6PD status in patients needs to be determined prior to prescription, and tafenoquine can be used only in patients having >70% of normal G6PD activity.⁶⁷ This stipulation is especially challenging in resource-limited countries. G6PD catalyzes the oxidation of glucose 6-phosphate to 6-phosphoglucono-δ-lactone, concomitant with NADPH generation in the cytosol.⁶⁸ Because human erythrocytes lack mitochondria, these cells heavily rely on G6PD-mediated NADPH production to provide sufficient reducing power against oxidative stress.^69,70 As such, G6PD-deficient erythrocytes have lower tolerability to oxidizing agents, including the hydroxylated metabolites of 8-aminoquinolines.⁷¹ Studies of G6PD deficiency and primaquine-induced hemolysis indicate that oxidation of hemoglobin and the membrane skeletal proteins, but not lipid peroxidation, leads to the membrane instability and RBC clearance.⁷²⁻⁷⁵ This protein oxidation event is accompanied by hemoglobin denaturation and formation of irreversible aggregates called the Heinz bodies. The occurrence of Heinz bodies is associated with the hemolytic toxicity of 8-aminoquinolines.⁷ However, the drug-induced oxidative stress may not be the only reason for Heinz body formation, given that treatment with different oxidants caused Heinz bodies to a varying degree.^7,76 The special relationship between 8-aminoquinoline and Heinz body formation could probably be recapitulated by oxidized phenylhydrazine that interacts with ferrihemoglobin to form ferrihemochrome.^7,77−80 Another player would be semiquinonimine radicals generated during redox cycling. These highly reactive intermediates may irreversibly modify hemoglobin via covalently attaching to nucleophilic residues and/or the porphyrin ring, thus aggravating hemoglobin precipitation. The ability to form adducts with the heme porphyrin might be analogous to that of endoperoxide antimalarials.^81,82 Biochemical analysis such as mass spectrometry could be useful for examining this hypothesis.⁸³ Overall, the exacerbated accumulation of ROS in G6PD-deficient patients and the molecular events involving tafenoquine metabolites may together damage the RBC membrane to eventually result in hemolytic anemia.

DRUG COMBINATIONS: BOOSTING THE CLINICAL UTILITY OF TAFENOQUINE

The synergism between 8-aminoquinolines and blood schizonticides has been noticed for decades with unknown mechanisms behind the drug−drug interactions.^7,31,84 In P. cynomolgi-infected monkeys, co-administration with chloroquine reduced the minimum curative dose of tafenoquine by 10-fold when compared to tafenoquine monotherapy.⁸⁵ Along tafenoquine’s clinical development path, it had been coupled with chloroquine, the standard of care for P. vivax infection, without the assessment of a possible synergistic effect with other blood schizonticides.^8,86 This drug combination is threatened by the establishment and spreading of chloroquine-resistant P. vivax.⁸⁷ Ideally, tafenoquine should be co-administered with fast-acting blood schizonticides such as artemisinin and its derivatives due to its slow parasite clearance rate.¹⁴ Current in vitro evidence suggests synergistic interactions between tafenoquine and artemisinin combination therapies (both artemisinin and its partner drugs) in P. falciparum blood-stage schizonts,^88,89 though if artemisinin (and other blood schizonticides) potentiates the radical curative activity or alleviates the toxicity of tafenoquine remains unclear.

The hemolytic toxicity in patients with G6PD deficiency has limited the clinical application of tafenoquine and other 8-aminoquinolines. Although high-throughput screening of detoxifying partner drugs has been proposed, such an assay has not been reported, partly due to the paucity of relevant lead compounds.⁷ A recent biochemical study identified a G6PD agonist AG1 that can activate wild-type and mutant G6PD enzymes.⁹⁰ AG1 treatment suppressed drug-induced oxidative stress in zebrafish and human RBCs and exhibited antihemolytic potential with enhanced G6PD activity.⁹⁰ Testing the ability of AG1 and other G6PD agonists to mitigate the drug-induced hemolysis without compromising the therapeutic activity of tafenoquine is a much-needed next step for the treatment of G6PD-deficient individuals.

NEXT-GENERATION 8-AMINOQUINOLINES

In search of novel 8-aminoquinolines with higher therapeutic indices to replace racemic primaquine, Schmidt and colleagues⁹¹ reported that (+)-primaquine was 3−5-fold less toxic than (−)-primaquine, while the capacities of primaquine and its enantiomeric forms to cure P. cynomolgi infection in a relevant non-human primate model remained the same. Later studies further revealed the enantiospecificity of primaquine with varied metabolic, pharmacodynamic, and pharmacokinetic behaviors in rhesus monkeys and humans.⁹²⁻⁹⁴ Thus, a path toward improved 8-aminoquinolines could be to evaluate the chemotherapeutic indices of individual enantiomers of promising candidates.

An 8-aminoquinoline analogue NPC1161C as a racemic form (also known as WR233078) exhibited superior blood schizonticidal activity with minimal toxicity in a P. berghei mouse model when compared to those of primaquine and tafenoquine (Figure 1).⁹⁵ This compound also showed excellent radical curative activity against P. cynomolgi infection in monkeys.⁹⁶ However, NPC1161C induced signs of hematotoxicity in beagles to a level higher than that induced by primaquine and tafenoquine.⁹⁵ Efforts to study the importance of stereochemistry in the antiparasitic and toxic activities of NPC1161C revealed that the (−)-enantiomer NPC1161B displayed reduced hematotoxicity and improved efficacy compared to those of its racemate.⁹⁵ NPC1161B has prophylactic activity in the P. berghei mouse model and is potent against P. cynomolgi liver-stage schizonts and hypnozoites. Additionally, this compound inhibits the P. falciparum sexual and asexual blood stages and sporozoite formation in Anopheles mosquitoes.^35,97,98 Whether the antirelapse and prophylactic activities of NPC1161B can be translated to human trials would depend on extensive pharmacological and safety profiles in animal models, especially in the setting of G6PD deficiency. To facilitate these steps, an established humanized mouse model with G6PD-deficient RBCs can be used to assess its hemolytic toxicity.⁹⁹

Given the intimate link between 8-aminoquinolines’ metabolism and their activities/toxicities, numerous primaquine derivatives have been synthesized through modification of the quinoline core and the terminal amino group to improve the therapeutic window.^100,101 Many primaquine derivatives showed reduced cytotoxicity in cell cultures in addition to higher potency against the P. falciparum blood stage and P. berghei liver stage in vitro.¹⁰¹ However, few studies have compared the efficacies of these derivatives in multiple Plasmodium life stages, and even fewer have assessed their radical curative activities and the hemolytic toxicity in the context of G6PD deficiency. As one example, a cohort of primaquine-thiazolidinones were generated and shown to suppress rodent and avian malaria transmission (Figure 1).¹⁰² Despite the slightly lower activity against P. berghei liver-stage schizonts in vitro and in vivo, the cytotoxicity and hemolytic toxicity were greatly reduced in different cell types as well as G6PD-deficient human RBCs.¹⁰² Perhaps similar modifications to tafenoquine or NPC1161B would facilitate the search for safer substitutes for treating G6PD-deficient individuals while retaining their long-acting, antirelapse, and broad-spectrum activities.

OUTLOOK

Since the introduction of the ancestral 8-aminoquinoline plasmochin in the early 1900s, the molecular mechanisms underlying the radical curative activity and the hemolytic toxicity of 8-aminoquinolines have been elusive.⁷ Technical challenges in studying the dormant liver-stage parasites, the inaccessibility of relevant G6PD-deficient animal models, and the difficult-to-identify metabolites due to their instability contribute to this lack of knowledge.^7,35 Undoubtedly, the characterization of parasites with resistance to 8-aminoquinolines would advance our mechanistic understanding, but these have yet to be isolated. When a more comprehensive understanding of the mechanisms of action of 8-aminoquinolines is achieved, it will accelerate next-generation antimalarial drug development and the strategic design of drug combinations. This will allow us to increase the therapeutic utility and antimalarial potency of these drugs, while minimizing the hemolytic toxicity in G6PD-deficient individuals. The ongoing development of P. cynomolgi and P. vivax culture systems,^35,45,103 the humanized G6PD-deficient mouse model,⁹⁹ and the advancing omics-based methods^16,17,26,104 should greatly aid in the understanding of such mechanisms.

The standard quantitative assessment of G6PD status is critical for the deployment of 8-aminoquinolines. However, such diagnostics are not easily accessible in resource-limited regions where malaria is endemic.^7,105,106 New drugs that can be safely administered to all patients will be crucial for malaria elimination, but the search for these new compounds was hampered by a lack of accessible tools. Recently, significant advances in model systems and screening tools, many spurred by support from The Bill & Melinda Gates Foundation (BMGF) and the Medicines for Malaria Venture (MMV), are accelerating the drug discovery process.³⁵ For example, Roth and colleagues¹⁰⁷ developed a platform to rapidly identify liver-stage schizonts and hypnozoites in P. vivax-infected primary human hepatocytes, which allows for automated drug screening with higher throughput. Important progress has also been made with the continuous culturing of blood-stage P. cynomolgi, which shares many biological features with P. vivax, including hypnozoite development.⁴⁵ While clearly an important tool for blood-stage P. cynomolgi drug screening, this long-term culture system may be further developed to promote gametocytogenesis and aid in liver-stage assays.⁴⁵ Additionally, the advent of humanized mouse models that support complete liver-stage and blood-stage parasite development will provide a more efficient preclinical assessment of candidate drugs.^108−110 These methodologies, among others, will accelerate the discovery and development of new lead compounds to facilitate malaria control efforts.

Like endoperoxides, drugs that induce general cellular stresses tend to be among the most effective and useful antimicrobial agents.^111,112 Multiple lines of evidence suggest that tafenoquine triggers oxidative stress (and proteotoxic stress) in Plasmodium parasites and G6PD-deficient RBCs.^{11,24,36,65,99} Although it is unclear whether tafenoquine kills Plasmodium at different life cycle stages in the same manner, tafenoquine’s activity against Leishmania and T. brucei appears to involve ROS production.^48,49 Through unknown mechanisms, tafenoquine also has activity against Toxoplasma gondii, Babesia microti, and Pneumocystis carinii.^113−115 It is likely that tafenoquine, once activated by CYP 2D enzymes or an alternative pathway, triggers general cellular stresses that then contribute to its broad-spectrum activity. With that being said, a fundamental understanding of how Plasmodium responds to various stresses is important. The front-line antimalarial artemisinin can generate oxidative stress and proteotoxic stress in Plasmodium and effectively kill the parasites.^38,39,116 Despite the exceptional potency of artemisinin, P. falciparum with increased tolerability to this drug has been detected in Southeast Asia.^117,118 Transcriptomic analyses have shown that genes involved in the unfolded protein response, oxidative stress response, and protein turnover are upregulated in artemisinin-resistant P. falciparum.^{38,39,119−121} As such, a systemic change in the Plasmodium transcriptional program renders a higher tolerability to drug-induced cellular stress. An in-depth understanding of tafenoquine’s mechanism of action and how Plasmodium reacts to the stress would be imperative. Inhibitors targeting the Plasmodium stress response, such as the parasite proteasome and the TCP-1 ring complex (TRiC), could eventually partner with 8-aminoquinolines in drug combinations to address parasites with high tolerability.

ABBREVIATIONS

RBC

red blood cell

CYP

cytochrome P450

ROS

reactive oxygen species

G6PD

glucose-6-phosphate dehydrogenase

NADPH

nicotinamide adenine dinucleotide phosphate

TQ

tafenoquine

PQ

primaquine