Antimalarial resistance risk in Mozambique detected by a novel quadruplex droplet digital PCR assay

Abstract

While the Plasmodium falciparum malaria parasite continues to cause severe disease globally, Mozambique is disproportionally represented in malaria case totals. Acquisition of copy number variations (CNVs) in the parasite genome contributes to antimalarial drug resistance through overexpression of drug targets. Of interest, piperaquine resistance is associated with plasmepsin 2 and 3 CNVs (pfpmp2 and pfpmp3, respectively), while CNVs in the multidrug efflux pump, multidrug resistance-1 (pfmdr1), increase resistance to amodiaquine and lumefantrine. These antimalarials are partner drugs in artemisinin combination therapies (ACTs) and therefore, CNV detection with accurate and efficient tools is necessary to track ACT resistance risk. Here, we evaluated ~300 clinically derived samples collected from three sites in Mozambique for resistance-associated CNVs. We developed a novel, medium-throughput, quadruplex droplet digital PCR (ddPCR) assay to simultaneously quantify the copy number of pfpmp3, pfpmp2, and pfmdr1 loci in these clinical samples. By using DNA from laboratory parasite lines, we show that this nanodroplet-based method is capable of detecting picogram levels of parasite DNA, which facilitates its application for low yield and human host-contaminated clinical surveillance samples. Following ddPCR and the application of quality control standards, we detected CNVs in 13 of 229 high-quality samples (prevalence of 5.7%). Overall, our study revealed a low number of resistance CNVs present in the parasite population across all three collection sites, including various combinations of pfmdr1, pfpmp2, and pfpmp3 CNVs. The potential for future ACT resistance across Mozambique emphasizes the need for continued molecular surveillance across the region.

Keywords: copy number variation; droplet digital PCR; drug resistance; malaria; multidrug resistance; plasmepsin.

Fig 1

Two-dimensional plot of the quadruplex ddPCR assay. Sixteen clusters of droplets (maximum possible) represent (1): droplets containing no target DNA (negative population), (2): droplets containing at least one copy of pfhsp70, (3): droplets containing at least one copy of pfmdr1, (4): droplets containing at least one copy of pfpmp3, (5): droplets containing at least one copy of pfpmp2, (6): droplets with both pfpmp3 and pfmdr1, (7): droplets with both pfpmp3 and pfhsp70, (8): droplets with both pfmdr1 and pfhsp70, (9): droplets with both pfhsp70 and pfpmp2, (10): droplets with pfmdr1, pfhsp70, and pfpmp3, (11): droplets with pfpmp2 and pfmdr1, (12): droplets with pfhsp70, pfmdr1, and pfpmp2, (13): droplets with pfpmp2 and pfpmp3, (14): droplets with pfhsp70, pfpmp2, and pfpmp3, (15): droplets with pfmdr1, pfpmp2, and pfpmp3, (16): and droplets with pfmdr1, pfpmp2, pfpmp3, and pfhsp70.

Fig 2

Parasite lines with known CNVs show the accuracy of the quadruplex ddPCR assay. All samples were run at 60 cycles. Clusters of droplets are colored as in Fig. 1. (A) Sample: laboratory-cultured P. falciparum NF54 DNA with no known CNVs in loci of interest (0.08 ng DNA, 17,739 total droplets, CN: 0.94/1.01/1.09; pfpmp3, pfpmp2, and mdr1, respectively). Note: this sample is also part of the dilution series represented in Fig S7. (B) Sample: laboratory-cultured P. falciparum Dd2 DNA with three copies of pfmdr1 (0.02 ng DNA, 18,400 total droplets, CN: 2.94; mdr1). (C) Sample: laboratory-cultured P. falciparum PM2GT clone F4 DNA with four copies of pfpmp2 (0.02 ng DNA, 9,690 total droplets, CN: 4.43; pfpmp2).

Fig 3

Summary of sample processing and results by province. (A) Depiction of Mozambique sites where samples were collected. Provinces are depicted in dashed lines. The city of collection within the province is depicted with a colored dot. (B) Sample quality control preformed from ddPCR data. Names of the province and city of collection (in parentheses) above pie charts representing the pass rate of quality control using λ (see Materials and Methods); the dashed portion represents failed quality control. n = total number of samples collected. (C) Results of ddPCR CNV analysis per-target in each province. n* = total number of samples passing quality screening in each province; 1.2–1.5 represents “potential CNVs”; >1.5 represents “called CNVs”; total CNVs are the sum of the CNVs > 1.2; %Province = Total CNVs/n*. (D) The cumulative total of all CNVs (>1.2) observed from all provinces. n* = combined total samples that passed quality screening.