Plasmodium vivax, a major cause of malaria in the Americas, Central and Southeast Asia, and Eastern parts of Africa, can result in severe and fatal illness . The P. vivax lifecycle includes a dormant liver stage, the hypnozoite, activation of which causes relapses. The WHO recommends treatment with a schizonticide to treat the blood stage, combined with hypnozoite clearance . Until recently, the only option available to kill hypnozoites was primaquine, first registered in the 1950s, and often associated with poor compliance because it requires administration once daily for 14 days . Tafenoquine is an 8-aminoquinoline with longer-acting antihypnozoite activity. Tafenoquine is a single-dose treatment, reducing the risk of noncompliance which is expected to result in improved individual and public health outcomes. Phase 2 and 3 trial results of recurrence prevention have consistently shown that tafenoquine given with chloroquine was superior to chloroquine alone [4,5] and is similar in efficacy and safety to primaquine given with chloroquine . Approximately, 15% of effective drugs are estimated to have robust genetic predictors of efficacy with roughly half of these being of clinical relevance . Therefore, GlaxoSmithKline routinely screens for genetic variants that may predict efficacy as part of clinical development. While there has been candidate gene research into chloroquine and primaquine efficacy in P. vivax treatment [8–10] including the effect of CYP2D6 on primaquine efficacy [8,10], this is the first exploration to examine genome-wide variation and tafenoquine treatment response.
Participants in this exploratory, retrospective pharmacogenetics study were enrolled in GATHER or DETECTIVE Parts 1 and 2 (ClinicalTrials.gov numbers NCT02216123 and NCT01376167, respectively). All patients were in the intent-to-treat (ITT) population and all but one patient, in DETECTIVE Part 1, had microscopically confirmed P. vivax microscopically confirmed ITT (mITT). Details on the studies can be found in the Supplementary material, Supplemental digital content 1, http://links.lww.com/FPC/B371. Participants were drawn from the DETECTIVE and GATHER mITT populations, provided written informed consent for genetics research, a DNA sample, and were successfully genotyped. All patients were treated with chloroquine on days 1–3 (600, 600, and 300 mg) to treat the blood stage of P. vivax malaria. The main pharmacogenetic (PGx) analysis group, PGx TQ, included patients on 300 or 600 mg of tafenoquine in combination with chloroquine; patients on 50 and 100 mg tafenoquine doses were not analyzed as these doses were found to not be efficacious . Two additional pharmacogenetic analysis groups were analyzed to aid in the interpretation of the PGx TQ results; the PGx CQ group consisting of patients treated with chloroquine alone and the PGx PQ group, patients treated with primaquine and chloroquine.
Definition of outcomes, genotyping and imputation details are summarized in the Supplementary materials, Supplemental digital content 1, http://links.lww.com/FPC/B371. Analysis was conducted using an additive genetic model on genome-wide variants with a minor allele frequency ≥0.01 and an imputation quality score ≥0.30. The 6- and 4-month recurrence-free efficacy outcomes were analyzed using a logistic regression model. Time to recurrence up to 6 months postdosing was analyzed using a Cox proportional hazards regression model. Models included an adjustment for region (South America, Asia, or Africa) and the first 10 genetic ancestry principal components. PGx TQ, PGx PQ, and PGx CQ included all three geographic regions, races, and ethnicities. Previously conducted population pharmacokinetic analysis for tafenoquine exposure across six clinical studies demonstrated that there was no clinically relevant impact of weight, BMI, age, or sex on tafenoquine area under the curve . Therefore, tafenoquine exposure is unlikely to be impacted by any of these factors and they were not included as covariates in the analysis models. The primary analysis was in PGx TQ where the power to detect genetic effects for recurrence-free outcomes exceeded 80% when variants had frequencies ≥10% and per allele odds ratios (ORs) were greater than four. Supplementary analyses were conducted in PGx PQ and PGx CQ and by the two largest geographic subgroups (South America and Asia) to aid in interpretation of the primary analysis results; the African subgroup was not analyzed separately due to its small sample size. The conventional P ≤ 5 × 10−8 threshold for declaring genome-wide significance for common variants was used . No multiplicity adjustment was made for multiple outcomes or analysis subgroups.
The pharmacogenetic sample was comprised of 900 patients of whom 529, 221, and 150 were treated with tafenoquine, primaquine, or chloroquine alone, respectively. From this, 73 of the 529 tafenoquine patients were excluded from analysis as they had been treated with 50 or 100 mg doses of tafenoquine. A further 44 patients were excluded due to close relatedness (third degree or closer) resulting in pharmacogenetic analysis groups of 438, 206, and 139 for tafenoquine, primaquine, and chloroquine, respectively. Demographic and efficacy information were similar across the pharmacogenetic analysis groups and their constituent mITT populations from which they were drawn (Table 1).
Genome-wide association results in the PGx TQ analysis group
The results for the three outcomes appear well calibrated as shown in the Manhattan and QQ plots (Supplementary Fig. 1a–f, Supplemental digital content 1, http://links.lww.com/FPC/B371). No significant association was observed for time to recurrence. Recurrence-free efficacy at 6 months was significantly associated with an intergenic variant, rs62103056 (hg19 position 18:41424097, imputation quality score of 1.0) where the A allele, with a frequency of 0.25, was associated with increased recurrence [OR = 2.98, 95% confidence interval (CI) 1.99–4.46, P = 3.79 × 10−8] (Fig. 1, Supplementary Fig. 2, Supplemental digital content 1, http://links.lww.com/FPC/B371). A similar trend was seen in the PGx TQ South American (OR = 3.39, 95% CI 2.07–5.56, P = 5.64 × 10−7) and PGx TQ Asian subgroups (OR = 3.10, 95% CI 1.33–7.25, P = 0.013). Results for recurrence-free efficacy at 4 months and time to recurrence show a similar trend in PGx TQ (Supplementary Fig. 3, Supplemental digital content 1, http://links.lww.com/FPC/B371). No significant association was seen for this variant in either PGx PQ (OR = 1.03, 95% CI 0.57–1.89, P = 0.92) or PGx CQ (OR = 0.46, 95% CI 0.20–1.05, P = 0.08). rs62103056 is approximately 12.5 kb 5′ of RNU6-443P (RNA, U6, small nuclear 443, pseudogene) (Supplementary Fig. 2, Supplemental digital content 1, http://links.lww.com/FPC/B371). A forest plot for the rs62103056 results for all three outcomes and all pharmacogenetic analysis groups is shown in Supplementary Fig. 4, Supplemental digital content 1, http://links.lww.com/FPC/B371.
Recurrence-free efficacy at 4 months revealed a significant signal in an 30-kb intergenic region on chromosome 12 (hg19 position 12: 89026155-89051634) where the minor alleles of six well imputed (imputation scores ≥ 0.96), highly correlated variants (R2: 0.87–0.94) with allele frequencies of ~2% were associated with improved efficacy. For purposes of illustration, results for rs11104986 will be discussed as this variant had the smallest P value. The minor G allele was associated with increased recurrence (OR = 42.30, 95% CI 7.95–225.10, P = 4.18 × 10−9) (Fig. 2, Supplementary Fig. 5, Supplemental digital content 1, http://links.lww.com/FPC/B371). The frequency of the G allele was ~2% in both the South American and Asian subgroups and the direction of effect was the same as in the overall PGx TQ group. Results for recurrence-free efficacy at 6 months and time to recurrence show a similar trend in PGx TQ, while no significant association was seen for this variant in either the PGx PQ (OR = 1.22, 95% CI 0.28–5.35, P = 0.81) or PGx CQ (OR = 0.57, 95% CI 0.06–5.59, P = 0.64) groups (Supplementary Fig. 6, Supplemental digital content 1, http://links.lww.com/FPC/B371). This region is approximately 22 and 52 kb 3′ of RNU1-117P and KITLG, respectively (Supplementary Fig. 5, Supplemental digital content 1, http://links.lww.com/FPC/B371). RNU1-117P is a small, nuclear pseudogene. KITLG (KIT ligand) is associated with the skin, hair, and eye pigmentation, and GTEx RNA-seq  expression data show the highest median expression in transformed fibroblasts.
Previous assessments of P. vivax drug response have focused on candidate gene host effects on chloroquine and primaquine treatment. This is the first pharmacogenetic study to assess the influence of host genome-wide variation on response to tafenoquine in the radical cure for P. vivax malaria and is powered at 80% to rule out large genetic effects with minor allele frequencies ≥10% and per-allele OR ≥ 4. Two signals passed the statistical threshold for genome-wide significance, but neither has a strong biological rationale after reviewing information from GTEx release 7 and the Ensembl Variant Effect Predictor CRCh37 release 95  and replication in an independent population is needed. The chromosome 12 signal, even if it were to replicate, would not provide much use in a clinical setting due to the low frequency of the minor allele. The chromosome 18 signal, rs62103056, if replicated, may be of clinical relevance in guiding treatment decisions as patients treated with tafenoquine and carrying the AA genotype show reduced efficacy; however, this replication would need to be done in a real-world setting that considers noncompliance with treatment.
In areas where malaria transmission is high, it is difficult to distinguish relapse, due to hypnozoite activation, from recurrence which includes hypnozoite activation and reinfection. A limitation of this study is that we cannot distinguish relapse from reinfection. Additional studies in areas of low or without malaria transmission would be needed to confirm that these results are related to relapse. This study is important in providing an initial view into the pharmacogenetics of tafenoquine efficacy in the treatment of P. vivax malaria. Further investigation into the role of host genetics and the interaction between pathogen and host may help elucidate the role of host genetics on tafenoquine efficacy.
We thank the patients and the following people for their contribution to this project: Cindy Chu [Mahidol University, Mae Sot, Thailand]; Mathias Chiano, Charlies Cox, Lynda Kellam, Joerg-Peter Kleim [GSK]; and Dana Fraser, Sandy Stinnett [Parexel].
Medicines for Malaria Venture (MMV) and GlaxoSmithKline (GSK) funded the main studies where DNA samples were collected. GSK funded the pharmacogenetic analyses (GSK identifier 208099) described in this article.
Anonymized individual participant data and study documents can be requested for further research from www.clinicalstudydatarequest.com
Conflicts of interest
The views expressed here are solely those of the authors and do not reflect the views, policies or positions of the U.S. Government or Department of Defense. Material has been reviewed by the Walter Reed Army Institute of Research. There is no objection to its presentation and/or publication. The opinions or assertions contained herein are the private views of the author, and are not to be construed as official, or as reflecting true views of the Department of the Army or the Department of Defense. The investigators have adhered to the policies for protection of human participants as prescribed in AR 70–25.
D.P. declares grants and nonfinancial support from GSK and MMV. A.L-C. declares grants and personal fees from GSK. D.Y. declares grants from GSK. J.B. and G.K. are GSK employees and hold shares/options in GSK. J.G. is a ViiV employee and holds shares/options in GSK. S.D. is an MMV employee. P.S. is an employee at Parexel whose work was funded by a contract with GSK. There are no conflicts of interest for the remaining authors.
1. Howes RE, Battle KE, Mendis KN, Smith DL, Cibulskis RE, Baird JK, et al. Global epidemiology of Plasmodium vivax
. Am J Trop Med Hyg. 2016; 95:15–34
2. World Health Organization. Guidelines for the Treatment of Malaria. 20153rd ed, Geneva: World Health Organization
3. Takeuchi R, Lawpoolsri S, Imwong M, Kobayashi J, Kaewkungwal J, Pukrittayakamee S, et al. Directly-observed therapy (DOT) for the radical 14-day primaquine treatment of Plasmodium vivax
malaria on the Thai-Myanmar border. Malar J. 2010; 9:308
4. Llanos-Cuentas A, Lacerda MV, Rueangweerayut R, Krudsood S, Gupta SK, Kochar SK, et al. Tafenoquine
plus chloroquine for the treatment and relapse prevention of Plasmodium vivax
malaria (DETECTIVE): a multicentre, double-blind, randomised, phase 2b dose-selection study. Lancet. 2014; 383:1049–1058
5. Lacerda MVG, Llanos-Cuentas A, Krudsood S, Lon C, Saunders DL, Mohammed R, et al. Single-dose tafenoquine
to prevent relapse of Plasmodium vivax
malaria. N Engl J Med. 2019; 380:215–228
6. Llanos-Cuentas A, Lacerda MVG, Hien TT, Vélez ID, Namaik-Larp RC, Chu CS, et al. Tafenoquine
versus primaquine to prevent relapse of Plasmodium vivax
malaria. N Engl J Med. 2019; 380:229–241
7. Nelson MR, Johnson T, Warren L, Hughes AR, Chissoe SL, Xu CF, et al. The genetics of drug efficacy: opportunities and challenges. Nat Rev Genet. 2016; 17:197–206
8. Marcsisin SR, Reichard G, Pybus BS.. Primaquine pharmacology in the context of CYP 2D6 pharmacogenomics: current state of the art. Pharmacol Ther. 2016; 161:1–10
9. Sortica VA, Lindenau JD, Cunha MG, O Ohnishi MD, R Ventura AM, Ribeiro-Dos-Santos ÂK, et al. SLCO1A2, SLCO1B1 and SLCO2B1 polymorphisms influences chloroquine and primaquine treatment in Plasmodium vivax
malaria. Pharmacogenomics. 2017; 18:1393–1400
10. Brasil LW, Rodrigues-Soares F, Santoro AB, Almeida ACG, Kühn A, Ramasawmy R, et al. CYP2D6 activity and the risk of recurrence of Plasmodium vivax
malaria in the Brazilian Amazon: a prospective cohort study. Malar J. 2018; 17:57
11. Thakkar N, Green JA, Koh GCKW, Duparc S, Tenero D, Goyal N.. Population pharmacokinetics of tafenoquine
, a novel antimalarial. Antimicrob Agents Chemother. 2018; 62:11
12. Dudbridge F, Gusnanto A.. Estimation of significance thresholds for genomewide association scans. Genet Epidemiol. 2008; 32:227–234
13. Gamazon ER, Segrè AV, van de Bunt M, Wen X, Xi HS, Hormozdiari F, et al.; GTEx Consortium. Using an atlas of gene regulation across 44 human tissues to inform complex disease- and trait-associated variation. Nat Genet. 2018; 50:956–967
14. Zerbino DR, Achuthan P, Akanni W, Amode MR, Barrell D, Bhai J, et al. Ensembl 2018. Nucleic Acids Res. 2018; 46:D754–D761