Liver-targeted polymeric prodrugs delivered subcutaneously improve tafenoquine therapeutic window for malaria radical cure

Abstract

Approximately 3.3 billion people live with the threat of Plasmodium vivax malaria. Infection can result in liver-localized hypnozoites, which when reactivated cause relapsing malaria. This work demonstrates that an enzyme-cleavable polymeric prodrug of tafenoquine addresses key requirements for a mass administration, eradication campaign: excellent subcutaneous bioavailability, complete parasite control after a single dose, improved therapeutic window compared to the parent oral drug, and low cost of goods sold (COGS) at less than $1.50 per dose. Liver targeting and subcutaneous dosing resulted in improved liver:plasma exposure profiles, with increased efficacy and reduced glucose 6-phosphate dehydrogenase-dependent hemotoxicity in validated preclinical models. A COGS and manufacturability analysis demonstrated global scalability, affordability, and the ability to redesign this fully synthetic polymeric prodrug specifically to increase global equity and access. Together, this polymer prodrug platform is a candidate for evaluation in human patients and shows potential for P. vivax eradication campaigns.

Fig. 1.. Conceptual overview of liver-targeted polymeric prodrug delivery mechanism.

The polymer prodrug with TQ is singly dosed SC and traffics in circulation to the liver. The polymeric prodrug achieves targeting and receptor-mediated endocytosis in hepatocytes via its GalNAc glycan ligand that is also incorporated into the polymer as a functionalized monomer. The blood-stabilized valine-citrulline linker is enzymatically cleaved by cathepsin enzymes in the endosomal-lysosomal compartments, and the TQ payload is released into the cytoplasm to kill dormant hypnozoites. Created with Biorender.com.

Fig. 2.. pSVCTQ results in optimal TQ release, polymer biodistribution specificity to liver, and subcellular hepatocyte localization.

(A) Liver and plasma concentrations of TQ after 25 mg/kg TQ equivalent dose of pSVCTQ or pVCTQ via IV route of administration. pSVCTQ dose results in lower plasma C_max compared to pVCTQ dose. (B) Liver and plasma concentrations of TQ after 10 mg/kg TQ equivalent dose of pSVCTQ via SC route of administration or 10 mg/kg oral dose of free TQ. pSVCTQ dose results in greater liver exposure and reduced plasma exposure compared to oral TQ. (C) IVIS imaging of relevant tissues at 8 hours post-SC 25 mg/kg dose of fluorescent rhodamine-labeled pSVCTQ. (D) Representative image of liver tissues prepared for imaging. Liver was collected at 8 hours post-SC 25 mg/kg dose of fluorescent rhodamine-labeled pSVCTQ. Five-micrometer-thick sections were prepared using CryoStat at −20°C, and the sections were counterstained with DAPI to visualize nuclei (blue) and Alexa 488 Phalloidin to visualize cell outlines (green). The polymer is in red. White arrows point to sinusoidal lumen. (E) VCTQ-MA containing PABC spacer and SVCTQ-MA without PABC spacer. All error bars shown are the SD from triplicate analyses. n = 3 per time point. LC-MS/MS technical replicates = 3. Monomer image and final figure created with Biorender.com.

Fig. 3.. pSVCTQ has approximately twice the activity of oral TQ in the prophylactic P. berghei model.

(A) Schematic of study design for the preexposure prophylactic model. Luciferase-expressing sporozoites are harvested from mosquitos. Mice are treated with either oral TQ or SC pSVCTQ and then inoculated with sporozoites. Bioluminescence of the sporozoites is recorded daily for 3 days using IVIS. From days 5 to 31, parasitemia is quantified using flow cytometry. (B) Results of the preexposure prophylactic dose-response study. Mice were dosed with either 10 mg/kg oral TQ or 5, 7.5, or 10 mg/kg TQ equivalent of pSVCTQ. No mice survived in the vehicle group. All mice survived in pSVCTQ 10 mg/kg SC group. n = 5 mice per treatment group. Schematic and final figure created with Biorender.com.

Fig. 4.. pSVCTQ reduces hemolytic toxicity in the humanized G6PD deficiency mouse model.

(A) Schematic depicting study design of in vivo hemolytic toxicity study. RBCs are collected from a G6PD-deficient donor (huRBCs) and then engrafted into NOD/SCID mice. Mice with >60% huRBCs at the end of the transfusion period (day 0) are dosed with varying concentrations of oral TQ or SC pSVCTQ polymer solution. (B) Dose-response comparison of oral TQ with SC pSVCTQ in the huRBC engrafted NOD/SCID hemolytic anemia model. pSVCTQ and TQ titration data are from a single experiment each. Data for 5 mg/kg oral TQ, which served as the positive control across studies, are from multiple experiments. The percent hemolysis was calculated for each treated animal from the percentage of huRBC present on days 0 and 7, followed by normalization to the vehicle control group. The normalized data were fit with a two-parameter “agonist versus normalized response” curve fit with shared Hill slope, using GraphPad Prism software. The doses resulting in 50% hemolysis (TD₅₀) and the 95% confidence intervals were 5.0 mg/kg (3.9 to 6.2) for oral TQ and 11.2 mg/kg (8.4 to 17.8) for pSVCTQ. An F-test analysis comparing the two data plots resulted in a P value for significance <0.0001. Vehicle, n = 3; all other groups, n = 4. Experiment was replicated three times. Oral TQ hist. indicates historical data for a hemotoxicity study using an oral dose of TQ. Schematic and final figure created with Biorender.com.

Fig. 5.. pSVCTQ shows anti-hypnozoite activity in the P. cynomolgi/rhesus primary liver cell model.

(A) Schematic of study design for anti-hypnozoite assay. Plate wells are seeded with hepatocytes on day −2. On hour 0, the cells are infected with P. cynomolgi sporozoites from Anopheles dirus salivary glands. Time is allowed for hypnozoite formation, and cells are treated with pSVCTQ solution on days 4 to 7. On day 8, the cells are fixed and stained for imaging and hypnozoite detection. (B) Anti-hypnozoite activity of pSVCTQ and toxicity to hepatocytes. (C) Free TQ is visualized inside cells when imaged at 355 nm following release from pSVCTQ. Cells were bathed in a solution of 500 μg polymer/ml PBS. Arrows indicate cell nucleus. (D) Representative images of ASGPR expression time course. Fixed cells were visualized using human ASGPR1 Alexa Fluor 647–conjugated antibody and then imaged at 647 nm. A steady increase of ASGPR is seen up to day 6 and then a slight decrease on day 8. The study was replicated twice. Schematic and final figure created with Biorender.com.

Fig. 6.. Engineering liver-targeted polymeric prodrug design for lower COGS.

Two designs were characterized and shown to reduce projected COGS while maintaining pharmaceutic functionality. In the first design, the copolymer was changed to a ter-copolymer design to substitute the more costly GalNAc targeting monomer with a less costly solubilizing monomer. The second design completely replaced the GalNAc targeting monomer, with the targeting ligand introduced via a tri-antennary GalNAc CTA. Created with Biorender.com.