Single-Dose Primaquine in a Preclinical Model of Glucose-6-Phosphate Dehydrogenase Deficiency: Implications for Use in Malaria Transmission-Blocking Programs

Abstract

Individuals with glucose-6-phosphate dehydrogenase (G6PD) deficiency (G6PDd) are at risk for developing hemolytic anemia when given the antimalarial drug primaquine (PQ). The WHO Evidence Review Group released a report suggesting that mass administration of a single dose of PQ at 0.25 mg of base/kg of body weight (mpk) (mouse equivalent of 3.125 mpk) could potentially reduce malaria transmission based on its gametocytocidal activity and could be safely administered to G6PD-deficient individuals, but there are limited safety data available confirming the optimum single dose of PQ. A single-dose administration of PQ was therefore assessed in our huRBC-SCID mouse model used to predict hemolytic toxicity with respect to G6PD deficiency. In this model, nonobese diabetic (NOD)/SCID mice are engrafted with human red blood cells (huRBC) from donors with the African or Mediterranean variant of G6PDd (A-G6PDd or Med-G6PDd, respectively) and demonstrate dose-dependent sensitivity to PQ. In mice engrafted with A-G6PD-deficient huRBC, single-dose PQ at 3.125, 6.25, or 12.5 mpk had no significant loss of huRBC compared to the vehicle control group. In contrast, in mice engrafted with Med-G6PDd huRBC, a single dose of PQ at 3.125, 6.25, or 12.5 mpk resulted in a significant, dose-dependent loss of huRBC compared to the value for the vehicle control group. Our data suggest that administration of a single low dose of 0.25 mpk of PQ could induce hemolytic anemia in Med-G6PDd individuals but that use of single-dose PQ at 0.25 mpk as a gametocytocidal drug to block transmission would be safe in areas where A-G6PDd predominates.

INTRODUCTION

Malaria is primarily a tropical disease with a global health burden of 1.2 billion people at high risk for infection, with approximately 438,000 deaths estimated in 2015 according to the World Health Organization (WHO) (1). Due to global eradication efforts, including the provision of insecticide-treated mosquito nets and artemisinin-based combination therapies (ACTs), malaria incidence has been on the decline for over a decade, with malaria deaths decreased by approximately 47% (1). Although these data are encouraging, effective drug therapy for reduction in transmission is a key factor in control programs and elimination tactics.

Primaquine (PQ) is the prototype 8-aminoquinoline (8-AQ) antimalarial drug and a potent gametocytocide in Plasmodium falciparum malaria (2). Despite its potential for use in malaria transmission-blocking programs (3) and ultimately eradication efforts (4), PQ has been underutilized and distribution has been severely restricted due to the risk of acute hemolytic anemia in populations that have glucose-6-phosphate dehydrogenase (G6PD) deficiency (G6PDd) (5, 6). G6PD deficiency is the most common human enzymopathy, with an estimated frequency of approximately 3 to 30% in areas where malaria is endemic (7–10). Coupled with the lack of available G6PD testing in the field, this puts large populations at possible risk for PQ-induced hemolysis if mass administration of PQ is done.

As recently reviewed by Steketee and ter Kuile (11), the WHO Evidence Review Group released a report in 2012 suggesting that a single dose of PQ at 0.25 mg of base/kg of body weight (mpk) could potentially reduce P. falciparum malaria transmission based on its ability to kill mature gametocytes and could safely be administered to individuals with P. falciparum infection, including those who are G6PD deficient, as a modification of the previous recommendation of 0.75 mpk of PQ by the WHO (12, 13). Unfortunately, there is limited safety evidence available confirming what the optimum single dose of PQ for treatment of P. falciparum malaria is in the context of G6PD deficiency.

Eziefula and colleagues conducted the first formal randomized control trial investigating the efficacy of single-dose PQ in a dose range of 0.1 mpk, 0.4 mpk, or 0.75 mpk combined with artemether-lumefantrine in a cohort of Ugandan children (14, 15). Based on a similar mean duration of gametocyte carriage, PQ at 0.4 mpk (6.3 days) or 0.75 mpk (6.6 days) showed similar gametocytocidal efficacies, with all tested doses well tolerated compared to the placebo. The authors noted an efficacy gradient from 0.4 mpk and 0.1 mpk of PQ that warrants further evaluation, especially with respect to the updated WHO recommendation of 0.25 mpk. However, this study was at the exclusion of G6PDd individuals, and only those with normal G6PD enzyme function as determined by fluorescent spot testing were enrolled. Concerns will linger over widespread deployment of PQ until further data are obtained to assess both its efficacy against P. falciparum gametocytes at relevant low doses and its safety in those with G6PD deficiency.

To address the safety of PQ at the dose newly recommended by the WHO, we designed a study for comparative assessment of the hemolytic potential of a single low-dose administration of PQ in a SCID mouse model using G6PDd red blood cells (RBC) from individuals of the African or Mediterranean genotype (A-G6PDd or Med-G6PDd, respectively), two major genetic variants of G6PD deficiency (16). Previously, we described the development and validation of a huRBC-SCID mouse model whereby nonobese diabetic (NOD)/SCID mice are given daily transfusions of human red blood cells (huRBC) from G6PDd donors (17). Our huRBC-SCID model is unique because it is the only readily available in vivo mouse model to test antimalarial drugs for their hemolytic toxicity in the background of G6PD deficiency. For this study, PQ was tested at doses either equivalent to or higher than the current recommendation of 0.25 mpk in A-G6PDd huRBC-SCID or Med-G6PDd huRBC-SCID mice to assess hemolytic toxicity and relative safety following treatment. We report here the first data showing that administration of the human-equivalent dose (HED) of 0.25 mpk of PQ does not cause hemolytic toxicity in A-G6PDd huRBC-SCID mice.

MATERIALS AND METHODS

huRBC-SCID mice engrafted with G6PD-deficient RBC.

The huRBC-SCID mouse model has been described elsewhere (17). Briefly, 5 × 10⁹ A- or Med-G6PDd human red blood cells (huRBC) in RPMI–25% autologous plasma were injected daily for 14 days into NOD.CB17-Prkdc^scid/J (NOD/SCID) mice aged 9 to 11 weeks (Jackson Laboratory, Bar Harbor, ME). Engraftment was assessed by flow cytometric analysis of the blood from engrafted mice using fluorescein isothiocyanate (FITC)-conjugated anti-glycophorin A monoclonal antibody for human RBC or FITC-conjugated CD71 and phycoerythrin (PE)-conjugated TER119 monoclonal antibody for murine reticulocytes and murine erythroid cell detection, respectively. Mice with peripheral huRBC levels greater than 60% were randomized for drug treatment, with 3 to 5 mice per group assigned. Tail snip blood was collected on days 0, 3, 5, and 7 of treatment for the determination of human red blood cell levels and murine reticulocyte levels. Percent loss of huRBC on day 7 was calculated as

$$\text{huRBC}\hspace{0.17em}\text{percent}\hspace{0.17em}\text{loss}=\left[\frac{(\text{day}\hspace{0.17em}0\hspace{0.17em}\text{value}-\text{day}\hspace{0.17em}7\hspace{0.17em}\text{value})}{\text{day}\hspace{0.17em}0}\right]\hspace{0.17em}\times \hspace{0.17em}100$$

Additionally, blood from G6PD-deficient volunteer blood donors was obtained for this study through the Walter Reed Army Institute of Research (WRAIR; Silver Spring, MD) in accordance with approved institutional review board protocols.

Drug treatment.

Drugs were provided by the Division of Experimental Therapeutics, WRAIR, and were reconstituted in the appropriate vehicle (phosphate-buffered saline [PBS]). Mice were treated with PQ once daily for 1 to 3 days per os (p.o.). On day 7 posttreatment, mice were sacrificed and assessed for percent huRBC, hematocrit levels and body, spleen, and liver weights. Liver and spleen weights were calculated as a percentage of total body weight for each mouse to account for variability in animal weights. All animals were treated according to SUNY Upstate Medical University's Committee for the Humane Use of Animals and in compliance with the Animal Welfare Act and other federal statutes and regulations relating to animals and experiments involving animals and adhere to the principles stated in the Guide for the Care and Use of Laboratory Animals (18).

Calculation of human-equivalent dose based on body surface area normalization.

The PQ dose translation from mouse to human, based on the U.S. Food and Drug Administration's body surface area (BSA) normalization method, was calculated using the following formula:

$$\text{HED}\hspace{0.17em}(in\hspace{0.17em}milligrams\hspace{0.17em}per\hspace{0.17em}ki\hspace{0.17em}\log ram)=animal\hspace{0.17em}dose\hspace{0.17em}(in\hspace{0.17em}milligrams\hspace{0.17em}per\hspace{0.17em}ki\log ram)\times \frac{animal\hspace{0.17em}{K}_m}{human\hspace{0.17em}{K}_m}$$

This calculation yielded a value for the human-equivalent dose (HED) and has been described previously (19, 20). In addition, the following average body weight, BSA, and K_m values for adult human and mouse were used (Table 1). K_m values are based on the average BSA calculations for humans and mice (K_m factor = body weight [in kilograms]/BSA [in square meters]). The K_m factor is not a true constant but is generally fixed for each species for standardization and practical purposes to convert a dose in milligrams per kilogram to a dose in milligrams per square meter. Each HED was calculated from PQ doses given to mice for this study (Table 2). For comparative purposes, the mouse equivalent dose of 9.25 mpk was also calculated based on the previous WHO recommendation of PQ at 0.75 mpk (HED).

TABLE 1.

Weight, BSA, and K_m values for adult human and mouse^a

Species	Wt (kg)	BSA (m2)	Km factor
Human	60	1.6	37
Mouse	0.02	0.007	3

^aThe K_m values were used to calculated human-equivalent dose.

TABLE 2.

Primaquine dose translation from mouse to human utilizing the body surface area normalization method^a

Mouse dose of PQ (mg of base/kg)	HED of PQ (mg of base/kg)
3.125	0.25
6.25	0.51
9.25	0.75
12.5	1
18.75	1.52
37.5	3.04

^aThe BSA normalization method was used to convert PQ treatment in mice to human-equivalent doses (HEDs) based on FDA guidelines as described in Materials and Methods. This range of PQ doses was selected for testing in G6PDd huRBC-SCID mice.

Pharmacokinetic analysis using dried blood spot cards.

Tail snip blood was taken at 45 and 120 min post-PQ dose, collected onto Whatman FTA DMPK-C filter paper cards, and allowed to dry. The FTA DMPK-C cards, Harris micropunch, Whatman 3.00-mm stainless steel tip, and Harris cutting mat were supplied by GE Healthcare (Piscataway, NJ). Sample tubes were obtained from Denville Scientific (Metuchen, NJ). Samples were extracted and analyzed as described below.

Sample preparation and extraction.

Standard curves and quality control (QC) samples were prepared by spiking mouse whole blood with PQ and carboxyprimaquine (CPQ) and then performing serial dilutions to desired concentrations. Aliquots of 15 μl of the spiked mouse whole blood from the standard curve and QC samples were spotted on Whatman FTA DMPK-C cards and were allowed to dry at 27°C. Two 3-mm cores were punched from the center of the 15-μl dried blood spot for extraction and placed into a clean Eppendorf tube. The 3-mm cores were submerged in 50 μl of methanol containing internal standard and vortexed gently for 15 min. The samples were then centrifuged at 16,000 × g for 15 min. The resulting supernatant was transferred to total recovery vials (Waters Corporation, Milford, MA) for liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis.

Instrumentation for PQ and CPQ analysis.

An AB Sciex 4000 Q-Trap triple quadrupole instrument (AB Sciex, Framingham, MA) with an electrospray interface in positive ionization mode was used for the multiple-reaction monitoring (MRM) LC-MS/MS analysis of all the samples. Electrospray ionization conditions were optimized for PQ and CPQ. Separations were achieved on a reversed-phase liquid chromatography Waters Xterra C₁₈ column (2.1 by 50 mm; 3.5 μm) under ambient (25 ± 2°C) conditions. The mobile phase consisted of solvent A, water with 0.1% formic acid, and solvent B, methanol with 0.1% formic acid. A 6.0-min gradient elution of the analytes was performed at a flow rate of 400 μl/min. PQ and CPQ concentrations were interpolated from a combined PQ and CPQ standard curve.

RESULTS

Pharmacokinetics of PQ and carboxyprimaquine in A-G6PDd huRBC-SCID mice.

Carboxyprimaquine (CPQ) is the predominant plasma metabolite of PQ (21, 22). Pharmacokinetic analysis of PQ was conducted to verify that PQ was efficiently metabolized to CPQ in the huRBC-SCID model. Blood samples were taken from A-G6PDd huRBC-SCID mice at 45 and 120 min following treatment with a single dose of PQ given orally at 6.25 or 12.5 mpk (Fig. 1). These sampling times correspond approximately to peak plasma levels at 45 min and near the end of the rapid decline phase after peak at 120 min for PQ and CPQ in mice (23, 24), demonstrating PQ conversion to CPQ.

FIG 1.

PK analysis of PQ-treated A-G6PDd huRBC-SCID mice. Shown are the results of a pharmacokinetic analysis of blood from A-G6PDd huRBC-SCID mice treated with 12.5 mpk or 6.25 mpk of PQ. The concentrations of PQ and CPQ detected in blood taken from mice at 45 min and 120 min after treatment with PQ at 12.5 and 6.25 mpk are shown. In mice treated with PQ at 12.5 mpk, 16.5 ng/ml of PQ and 160.8 ng/ml of CPQ were detected at 45 min posttreatment (SD = 8.5 and 74.1, respectively); PQ at 13.8 ng/ml and CPQ at 128.0 ng/ml were detected at 120 min posttreatment (SD = 4.4 and 31.2, respectively). In mice treated with PQ at 6.25 mpk, 11.3 ng/ml of PQ and 169.6 ng/ml of CPQ were detected at 45 min posttreatment (SD = 2.0 and 12.0, respectively); 8.9 ng/ml of PQ and 74.0 ng/ml of CPQ were detected at 120 min posttreatment (SD = 2.5 and 12.7, respectively). Each sample was analyzed twice (n = 4 of 5 mice per group). Values are means ± SD.

PQ dose translation from mouse to human utilizing the BSA normalization method.

The body surface area (BSA) normalization method was used to convert PQ treatment in mice to human-equivalent doses (HEDs) utilizing the FDA's recommendation for normalizing doses in accordance with body surface area as described above (see Materials and Methods). The HED for an adult human was calculated as the animal dose (in milligrams per kilogram) multiplied by the ratio of animal K_m to human K_m, where the K_m factor was based on the average BSA calculations for human and mouse for this study (adapted from the work of Reagan-Shaw et al. [19] and current FDA guidelines [20]; see Materials and Methods). The range of PQ doses selected for testing in huRBC-SCID mice in this study and their corresponding HEDs are outlined in Materials and Methods. The HED of 0.75 mpk (9.25 mpk in mice) shown in Table 2 demonstrates the difference between the previously recommended 0.75-mpk PQ single-dose regimen versus the updated recommendation of a 0.25-mpk PQ single-dose regimen by the WHO.

huRBC engraftment levels and huRBC loss in PQ treated A- and Med-G6PDd huRBC-SCID mice.

To determine the hemolytic potential of single-dose PQ with respect to individuals who are G6PD deficient, we tested a single dose of PQ in NOD/SCID mice engrafted with RBC from a G6PD-deficient donor with the A or Med variant. A-G6PDd huRBC-SCID mice were treated with a vehicle only (PBS), single doses of PQ at 3.125, 6.25, and 12.5 mpk, and 37.5 mpk of PQ (administered as 12.5 mpk/day for 3 days). We have previously shown the last dose to induce significant loss of huRBC in this model, indicative of hemolytic toxicity (17). Mice were then assessed for huRBC levels following PQ treatment on days 0, 3, 5, and 7 posttreatment. In mice treated with a single dose of PQ at 3.125, 6.25, or 12.5 mpk, the percentage of huRBC over 7 days was similar to that in the vehicle control group (Fig. 2A). However, by day 7 posttreatment, A-G6PDd huRBC-SCID mice treated with 12.5 mpk of PQ for 3 days (37.5 mpk of PQ total) had significantly reduced percentages of huRBC compared to those of the vehicle control and all treatment groups, indicative of hemolytic toxicity.

FIG 2.

PQ single-dose treatment in A-G6PDd huRBC-SCID mice. (A) Assessment of huRBC engraftment levels on days 0, 3, 5, and 7 after initiation of PQ treatment, shown as a percentage of total RBC in peripheral blood. Mice were treated with a vehicle only (PBS), single-dose PQ at 3.125 mpk, single-dose PQ at 6.25 mpk, single-dose PQ at 12.5 mpk, or PQ at 37.5 mpk (12.5 mpk/day for 3 days). (B) Percentage of murine reticulocytes and spleen weight normalized relative to body weight were assessed at 7 days posttreatment as indicators of hemolytic toxicity. (C) Percent loss of huRBC on day 7 posttreatment in A-G6PDd huRBC-SCID mice. For comparison, the 0.25-mpk human-equivalent dose of PQ is also shown. All statistically significant differences on day 7 were determined using a one-way analysis of variance (ANOVA) with Bonferroni posttest. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (n = 3 or 4 mice per group).

Consistent with a lower level of G6PD, our previous studies indicated that PQ given at 12.5 mpk/day for 7 days induces more significant hemolysis in Med-G6PDd huRBC-SCID mice than in A-G6PDd huRBC-SCID mice (17). Med-G6PDd huRBC-SCID mice were treated with a vehicle only (PBS), 3.125, 6.25, and 12.5 mpk of PQ given as a single dose, and 18.75 mpk of PQ (administered as 6.25 mpk/day for 3 days) and 37.5 mpk of PQ (administered as 12.5 mpk/day for 3 days). In Med-G6PDd huRBC-SCID mice treated with a single dose of PQ at 3.125, 6.25, or 12.5 mpk, a more rapid loss of huRBC was observed than in mice given the vehicle only. Med-G6PDd huRBC-SCID mice given PQ for 3 days at 12.5 mpk/day had an almost complete loss of huRBC by 5 days after initiation of treatment, while the kinetics of huRBC percentages in Med-G6PDd huRBC-SCID mice given the lower dose of 6.25 mpk (for 3 days) more closely mirrored the loss of huRBC in mice given a single dose of PQ (Fig. 3A).

FIG 3.

PQ single-dose treatment in Med-G6PDd huRBC-SCID mice. (A) Assessment of huRBC engraftment levels on days 0, 3, 5, and 7 after initiation of PQ treatment, shown as a percentage of total RBC in peripheral blood. Mice were treated with a vehicle only (PBS), single-dose PQ at 3.125 mpk, single-dose PQ at 6.25 mpk, single-dose PQ at 12.5 mpk, PQ at 18.75 mpk (PQ dosed at 6.25 mpk/day for 3 days), or PQ at 37.5 mpk (PQ dosed at 12.5 mpk/day for 3 days). (B) Percentage of murine reticulocytes and spleen weight normalized relative to body weight were assessed at 7 days posttreatment as indicators of hemolytic toxicity. (C) Percent loss of huRBC on day 7 posttreatment in Med-G6PDd huRBC-SCID mice. For comparison, the 0.25-mpk PQ human-equivalent dose is also shown. All statistically significant differences on day 7 were determined using one-way ANOVA with Bonferroni posttest. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 (n = 3 or 4 mice per group).

Other markers of hemolytic toxicity, including percentage of murine reticulocytes and spleen weight, were also assessed on day 7 posttreatment in both A- and Med-G6PDd huRBC-SCID mice. In A-G6PDd huRBC-SCID mice, findings were consistent with PQ-induced hemolysis when PQ was given for 3 days at 12.5 mpk/day, as indicated by elevated reticulocyte levels and spleen weight (Fig. 2B). Elevated reticulocyte levels and increased spleen weight were also apparent in Med-G6PDd huRBC-SCID mice at day 7 when PQ was given for 3 days at 12.5 mpk/day (Fig. 3B).

The percent huRBC loss was determined on day 7 post-PQ treatment in both groups of mice. A-G6PDd huRBC-SCID mice treated with a single dose of PQ at 3.125, 6.25, or 12.5 mpk did not show significant loss of huRBC at 7 days posttreatment compared to the vehicle control (Fig. 2C). In contrast, mice treated with 37.5 mpk of PQ (PQ dosed at 12.5 mpk/day for 3 days) showed a significant loss of huRBC at day 7 compared to the vehicle control (Fig. 2C). The WHO Evidence Review Group's recommendation is a single treatment at 0.25 mpk of PQ HED (equivalent to 3.125 mpk of PQ in mice), which does not induce hemolytic toxicity in A-G6PDd huRBC-SCID mice. However, evaluation of huRBC loss in Med-G6PDd huRBC-SCID mice indicated that even at the lowest dose of PQ, 3.125 mpk (the equivalent of the human recommended dose), the loss of huRBC was significantly greater than in mice given the vehicle control (Fig. 3C). This observation suggests that if the current recommended dose of PQ (0.25 mpk) for transmission-blocking programs was used, persons with Med-G6PDd could be susceptible to hemolytic toxicity.

DISCUSSION

In light of the recently updated WHO recommendation (12) for widespread deployment of a single dose of PQ at 0.25 mpk, a more refined safety profile for PQ as a P. falciparum gametocytocide with respect to G6PD deficiency is critical. Recent clinical trials for assessment of single low-dose PQ treatment of malaria have pointedly excluded individuals with G6PD deficiency; therefore, the potential for hemolytic toxicity in individuals with G6PD deficiency remains largely unknown (14, 25). We have developed a validated animal model to assess hemolytic toxicity (17). In this model, immunodeficient mice are transfused with huRBC from A- or Med-G6PDd huRBC. Treatment with PQ and other 8-AQs known to cause hemolytic toxicity induce a dose-dependent loss of huRBC that is commensurate with the level of G6PD deficiency (5, 6). In the present study, we evaluated whether a single dose of PQ given at the HED of 0.25 mpk would induce hemolytic toxicity in either A- or Med-G6PDd huRBC-SCID mice. We found that treatment of A-G6PDd huRBC-SCID mice with a single dose of PQ at an HED of 0.25 mpk showed no significant loss of huRBC compared to the value for vehicle control mice, whereas the Med-G6PDd huRBC-SCID mice treated at an HED of 0.25 mpk showed significant huRBC loss on day 7 post-PQ treatment compared to the vehicle control. Our data suggest that use of single-dose PQ at 0.25 mpk to block transmission may be safe in areas where the A-G6PDd variant is prevalent.

A major consideration in the administration of single low-dose PQ is the geographical region in which it will be deployed. In sub-Saharan Africa, where P. falciparum predominates, a single low dose of PQ could aid in the treatment of malaria in elimination programs. The A-G6PDd variant also predominates in sub-Saharan Africa, with approximately 90% of G6PD-deficient individuals have the A variant (8, 9). In countries such as Mali and Nigeria, where the A variant accounted for 100% of screened populations in large case control studies, single-dose PQ could potentially be distributed without G6PD testing and without posing a significant risk to G6PDd individuals (26, 27).

Since the Med variant is primarily found in populations within the Mediterranean region, including people of Italian, Grecian, Spanish, Arabic, and Jewish (Kurdish) descent, in addition to subset populations within parts of Malaysia, India, the Middle East, and the Comoros (26, 28), our data suggest that mass administration of PQ would be unsuited to these areas and its use would necessitate G6PDd testing first. A study published by Kheng et al. in 2015 assessed the tolerability of PQ in Plasmodium vivax-infected G6PDd individuals primarily of the Viangchan variant, a class II G6PDd variant marked by a more severe G6PD deficiency similar to Med-G6PDd (29). PQ given at approximately 0.75 mpk weekly beginning on day 0 (once per week for 8 weeks) with dihydroartemisinin-piperaquine on days 0, 1, and 2 (once daily for 3 days) was generally well tolerated, with no severe anemia observed, but the authors acknowledged that a quarter of the G6PDd patients experienced a >25% drop in hemoglobin concentrations, with nearly one-third of patients exhibiting PQ toxicity by day 7 posttreatment and one patient even requiring a blood transfusion (29). Therefore, they concluded that their data do not support unsupervised weekly PQ administration to prevent relapse in areas with severe G6PD deficiency and recommended G6PDd testing beforehand, though they speculated safety in mass administration of a lower (0.25 mpk) dose to block P. falciparum transmission in the context of circumventing the spread of artemisinin resistance that has emerged in the Greater Mekong Subregion, including the Cambodian-Thai and Thai-Myanmar borders in Southeast Asia (30). Although we did not test an HED of 0.75 mpk, we did test a higher dose with an HED of 1 mpk and a lower dose with an HED of 0.51 mpk in Med-G6PDd huRBC-SCID mice and found a significant loss of huRBC in both groups following PQ treatment compared to vehicle treatment, suggesting that PQ at 0.75 mpk has the potential to have hemolytic activity in a severely G6PD-deficient individual. Furthermore, while PQ treatment at an HED of 0.25 mpk showed less drastic indicators of hemolytic toxicity than did the higher doses in Med-G6PDd huRBC-SCID mice, potentially leaving open the question of human tolerability, we remain disinclined to suggest administration of PQ at any dose to severely G6PD-deficient individuals without further studies more clearly defining dose versus risk and the likely outcome of treatment.

We recognize that until we fully understand what species are responsible for efficacy and toxicity in humans following PQ treatment (22) and therefore understand the differential drug metabolisms and pharmacokinetics between mice and humans, we are unable to predict with complete certainty that the dose chosen in mice would be relevant from a human translational perspective until further studies are conducted. While our findings are indeed based on data extrapolated and interpreted from an animal study rather than a clinical trial, we have been able to demonstrate dose-dependent loss of G6PD-deficient RBC following PQ treatment in our SCID model, reproducing drug-induced hemolytic toxicity (17). This study was a necessary means to avoid the risks and ethical concerns involved in conducting an initial “trial and error” safety assessment of a hemolytic drug in G6PD-deficient individuals, and it is a good indicator for the potential tolerability of low-dose PQ in people with the enzyme deficiency. Accordingly, we chose G6PD-deficient donors of the A and the Med variants, which represent the two extremes of the G6PD spectrum, with the A variant phenotype representing a milder enzyme deficiency than the severely deficient Med variant (26). In consideration of the lack of extensive safety data available, we feel that this study provides compelling and important safety data that support PQ single-dose administration without G6PDd testing in geographical areas primarily represented by populations that are A-G6PD deficient. Still, because of the hemolysis observed in the Med-G6PDd huRBC-engrafted mice, we recommend exercising substantial caution and conducting G6PDd testing in areas where the Med variant predominates.

In 2014, Chen et al. (31) reported that seven African countries, including Botswana, Ethiopia, Namibia, South Africa, Swaziland, Zanzibar, and Zimbabwe, had already incorporated single low-dose PQ at 0.25 mg base per kg into their policy documents. However, this group followed up with a study in 2015 identifying major supply, programmatic, and regulatory barriers to the rollout of single low-dose PQ in some of these countries in addition to Senegal and Zambia (31, 32). Interviews conducted with key informants covering issues from PQ supply to dosing and to country perspectives showed that the regulatory processes at the country and global levels as well as lingering concerns over safety in G6PD-deficient individuals served as barriers to implementation in actual field settings. While the policies exist, hesitation over use will likely remain until convincing evidence is provided to support the widespread use of single-dose PQ.

We designed this study to augment the limited safety data showing single-dose PQ treatment with respect to G6PD deficiency and the new recommendation by the WHO Evidence Review Group. Overall, our data suggest that single-dose PQ at 0.25 mpk to block transmission would be safe in areas where the A-G6PDd variant is prevalent. Additionally, in terms of administration, we recommend caution in areas where the Med-G6PDd variant or other variants of severe G6PD deficiency predominate. Further studies defining the efficacious profile for the 0.1-mpk to 0.4-mpk efficacy gradient (14), including the recommended 0.25-mpk dose, as well as studies in a variety of settings where malaria is endemic will be needed to confirm the recommended dose with respect to G6PD deficiency.

Funding Statement

The National Center for Natural Products Research (NCNPR) at the University of Mississippi is also supported by USDA-ARS scientific cooperative agreement no. 58-6408-2-0009.