Genomic Analyses Reveal the Common Occurrence and Complexity of Plasmodium vivax Relapses in Cambodia

Abstract

Plasmodium vivax parasites have a unique dormant stage that can cause relapses weeks or months after the initial infection. These dormant parasites are among the main challenges of vivax malaria control as they constitute a reservoir that is difficult to eliminate. Since field studies are confounded by reinfections and possible recrudescence of drug-resistant parasites, most analyses of P. vivax relapses have focused on travelers returning from regions of malaria endemicity. However, it is not clear whether these individuals accurately recapitulate the relapse patterns of repeatedly infected individuals residing in areas of endemicity. Here, we present analyses of vivax malaria patients enrolled in a tightly controlled field study in Cambodia. After antimalarial drug treatment was administered, we relocated 20 individuals to a nontransmission area and followed them for 60 days, with blood collection performed every second day. Our analyses reveal that 60% of the patients relapsed during the monitoring period. Using whole-genome sequencing and high-throughput genotyping, we showed that relapses in Cambodia are often polyclonal and that the relapsing parasites harbor various degrees of relatedness to the parasites present in the initial infection. Our analyses also showed that clone populations differed dynamically, with new clones emerging during the course of the relapsing infections. Overall, our study data show that it is possible to investigate the patterns, dynamics, and diversity of P. vivax relapses of individuals living in a region of malaria endemicity and reveal that P. vivax relapses are much more pervasive and complex than previously considered. (This study has been registered at ClinicalTrials.gov under registration no. NCT02118090)IMPORTANCEP. vivax parasites can remain dormant in the liver and relapse weeks or months after the initial infection, greatly complicating malaria control and elimination efforts. The few investigations of this dormant stage have relied on travelers and military personnel returning from areas of malaria endemicity. However, it is not clear whether these individuals, exposed to a limited number of infections, accurately represent the patterns of relapses of individuals living in areas of endemicity, who are repeatedly infected by P. vivax parasites. Our study combined tightly controlled fieldwork with comprehensive genomic analyses, and our report provides a first opportunity to investigate the patterns, dynamics, and diversity of P. vivax relapses directly with individuals living in areas of endemicity.

Keywords: Plasmodium vivax; genomics; malaria; relapse.

Copyright © 2018 Popovici et al.

Figures

FIG 1

P. vivax positivity following chloroquine treatment. The figure shows the results of P. vivax detection following treatment (at day 0) for the patients (displayed as rows) for the 60-day monitoring period (x axis). Dark red indicates P. vivax-positive samples detected by both microscopy and qPCR, while light red indicates samples positive only by qPCR and green denotes P. vivax-negative samples. The yellow boxes indicate cases where a patient developed malaria symptoms and was retreated with chloroquine. "W" and "G" indicate instances where a blood samples was successfully characterized by P. vivax whole-genome sequencing (coverage of >50×) and genotyping (>50 SNPs), respectively. Black boxes indicate missing data.

FIG 2

Genetic relationships among the dominant clones of each infection. The figure shows a neighbor-joining tree reconstructed using the number of nucleotide differences between the genomes of each clone. The parasites dominant in the recurring infections are highlighted with colored circles and the clones dominant in the corresponding primary infection with triangles of the same color. (The initial infection of BL08 is not included as the sequencing data did not fulfill our quality control [QC] criteria).

FIG 3

Relatedness between relapsing parasites and those present in the initial infection. The figure shows the distribution, throughout the genome, of the new alleles (corresponding to patients BL010 [A], BL012 [B], BL002 [C], BL018 [D], and BL011 [E]) that were detected in the relapse infection but not observed in the initial infection. Each cross represents a 10-kb window containing more than five new alleles and is displayed according to its genomic position (x axis, nucleotide position in kilobases; y axis, chromosome). For patient BL10, the new alleles are distributed throughout the genome, suggesting that the initial and relapsing parasites are unrelated (A), while for the other patients, some large chromosomic regions (e.g., most of chromosome 12 in BL12 [B]) are free of new alleles, indicating identical genetic makeup characteristics and close relatedness among the initial and relapsing clones.

FIG 4

Complex changes in parasite populations detected by whole-genome sequencing. The figure shows the reference allele frequency plots generated from the parasites present in BL02 at days 41, 48, and 55 (left panel). Note the multiple modes at day 41 and day 55, indicating the presence of three main clones, while the lack of distinguishable modes at day 48 suggests the presence of either a single clone or numerous genetically different parasites in various proportions. These variations in the numbers of clones that were detected during the course of the infection are also reflected in the genotyping (gtyping) data generated from the same individual (gray curve, right panel) and accompany dramatic changes in parasitemia as determined by microscopy (orange curve) and qPCR (blue curve).

FIG 5

Schematic model of P. vivax relapses. (A) Following the bite of an infected mosquito, sporozoites from genetically different P. vivax parasites invade hepatocytes, undergo schizogony, and are released in the bloodstream and cause symptoms. A subset of the sporozoites becomes hypnozoites and remain dormant, together with hypnozoites from previous infections (in yellow). (B and C) These hypnozoites then reactivate later, either as an effect of an external stimulus (B) or stochastically (C). Our findings favor the idea of continuous, stochastic reactivation of hypnozoites from the liver (C) as our study did not reveal any stimulus whereas we observed new clones emerging during the course of the relapses.