Baseline Ex Vivo and Molecular Responses of Plasmodium falciparum Isolates to Piperaquine before Implementation of Dihydroartemisinin-Piperaquine in Senegal

Dihydroartemisinin-piperaquine, which was registered in 2017 in Senegal, is not currently used as the first-line treatment against uncomplicated malaria. A total of 6.6% to 17.1% of P. falciparum isolates collected in Dakar in 2013 to 2015 showed ex vivo-reduced susceptibility to piperaquine.

ABSTRACT

Dihydroartemisinin-piperaquine, which was registered in 2017 in Senegal, is not currently used as the first-line treatment against uncomplicated malaria. A total of 6.6% to 17.1% of P. falciparum isolates collected in Dakar in 2013 to 2015 showed ex vivo-reduced susceptibility to piperaquine. Neither the exonuclease E415G mutation nor the copy number variation of the plasmepsin II gene (Pfpm2), associated with piperaquine resistance in Cambodia, was detected in Senegalese parasites.

INTRODUCTION

Since 2005, the World Health Organization (WHO) has recommended artemisinin-based combination therapy (ACT) as the first-line treatment against malaria. The emergence of resistance to artemisinins, manifested by delayed parasite clearance after monotherapy with artesunate or ACT, was recently described in Southeast Asia (1, 2). Dihydroartemisinin-piperaquine is one of the most recent ACTs to be commercialized. However, multidrug resistance to dihydroartemisinin-piperaquine is currently emerging in Cambodia, where recrudescent infections have rapidly increased (3–5). Additionally, in vitro resistance to piperaquine was detected and increased rapidly between 2013 and 2015 in Cambodia (6).

Very few cases of resistance to dihydroartemisinin-piperaquine have been described in Africa (7). Two cases of late treatment failure after 30 and 32 days was reported in Italian travelers returning from Ethiopia and who were treated with dihydroartemisinin-piperaquine (8, 9).

Dihydroartemisinin-piperaquine obtained marketing authorization in June 2017 in Senegal. Artemether-lumefantrine and artesunate-amodiaquine combinations are currently the ACTs that are most used in Senegal. Dihydroartemisinin-piperaquine demonstrated a high efficacy in the treatment of uncomplicated malaria or in seasonal intermittent preventive treatment in children in Senegal (10–13).

Recently, an increase in the copy number of the plasmepsin II gene (Pfpm2) (PF3D7_1408000), which encodes a protease involved in hemoglobin degradation, has been associated with dihydroartemisinin-piperaquine treatment failures and raised piperaquine 50% inhibitory concentrations (IC₅₀s) in Cambodia (14, 15). The presence of parasites carrying 2 copies of Pfpm2 was confirmed in 7 infections out of 65 clinical failures to dihydroartemisinin-piperaquine in Mali (16).

The exonuclease gene (PF3D7_1362500) was one of the 40 candidate drug targets for Plasmodium falciparum identified by in silico prediction (17). The mutation E415G in the exonuclease gene (exo-E415G) was correlated with the high IC₅₀ of piperaquine in 297 isolates from Cambodia (Pursat, Preah, and Ratanakiri) obtained from 2011 to 2013 (15). This exo-E415G mutation was associated with parasite recrudescence following dihydroartemisinin-piperaquine treatment.

The aim of this study was to establish a susceptibility baseline of P. falciparum parasites to piperaquine before the implementation of the dihydroartemisinin-piperaquine combination in Senegal by evaluating P. falciparum parasite ex vivo susceptibility to piperaquine and the prevalence of the molecular markers potentially involved in piperaquine resistance.

A total of 76 P. falciparum clinical isolates collected by venous blood samples before treatment from falciparum malaria patients treated at the Hôpital Principal de Dakar from November 2013 to January 2014, August to December 2014, and September to November 2015 were assessed for ex vivo susceptibility to piperaquine and evaluated for the copy number of Pfpm2 and the exo-E415G mutation. Seventy-five percent of the patients were recruited from the emergency department. The patients were treated with quinine until 2014 and then with artesunate, artemether-lumefantrine, or artesunate-amodiaquine.

Thin blood smears were stained using a RAL kit (Réactifs RAL, Paris, France) based on eosin and methylene blue and were examined to determine the infection by P. falciparum and parasitemia. Parasitemia ranged from 0.06% to 14.4%. The diagnosis of monospecies infection by P. falciparum was confirmed by real-time PCR (LightCycler 2.0; Roche Group, Basel, Switzerland), as previously described (18).

The susceptibility of the isolates was assessed without culture adaptation. The parasitized red blood cells (final maximum parasitemia, 0.5%; final hematocrit, 1.5%) were incubated in 96-well plates predosed with antimalarial drugs (piperaquine [PPQ], chloroquine [CQ], quinine [QN], monodesethylamodiaquine [DQ], mefloquine [MQ], lumefantrine [LMF], and dihydroartemisinin [DHA]) for 72 h at 37°C with atmospheric generators for capnophilic bacteria using Genbag CO2 at 5% CO₂ and 15% of O₂ (bioMérieux, Marcy l'Etoile, France), as previously described (19, 20). PPQ was first dissolved in methanol and then diluted in water to final concentrations that ranged from 1.9 to 998 nM. An ex vivo histidine-rich protein 2 (HRP2)-based enzyme-linked immunosorbent assay (ELISA) was then performed using a commercial Malaria Ag competitive enzyme-linked immunosorbent assay (CELISA) kit (reference no., KM2159; Cellabs PTY Ltd., Brookvale, Australia), as previously described (19, 20). Each batch of 96-well plates predosed with antimalarial drugs was validated on the CQ-resistant W2 strain (isolated in Indochina; obtained from MR4, VA, USA) in four independent experiments using the same conditions described below. The mean 50% effective concentration (EC₅₀) value (effective concentration of a drug that prevents 50% of the growth observed in an untreated control) for the W2 strain for the different batches used during the 3 years was 32.5 nM for PPQ (and 292 nM for CQ, 275 nM for QN, 72 nM for DQ, 13.7 nM for LMF, 15.4 nM for MQ, and 1.27 nM for DHA). A comparison of W2 susceptibility data for the 7 antimalarial drugs indicated that there was no significant difference in the responses to antimalarial drugs over the 3 years (0.39 < P < 0.95). Additionally, the polymorphic genetic markers msp1 and msp2 and microsatellite markers specific to P. falciparum were genotyped at least once a month to verify W2 clonality (21, 22).

The copy number of Pfpm2 (PF3D7_140800) was assessed by quantitative PCR on the LightCycler 2.0 (Roche, Germany), as previously described (23). Isolates with a copy number greater than 1.6 were classified as isolates with 2 copies (14). An isolate collected from a patient with imported malaria after returning from Cambodia and harboring 2 copies of Pfpm2 was used as positive control (sample kindly provided by S. Houzé).

The two exons of the exonuclease gene (PF3D7_1362500) were amplified by PCR using the following primer pairs: Exo1_F, 5′-GGAACATAAGAAGGGTACTGAGC-3′ and Exo1_R, 5′-ATCTGATGATAGGAGGTGCAA-3′ for exon 1; and Exo2_F, 5′-GAATGGAGTCATTTAGCAGCAA-3′ and Exo2_R, 5′-CATTCCCATGTACTATCTTTGAACTT-3′ for exon 2. For each exon, a standard PCR was carried out with a 25-μl total reaction mixture containing 200 ng of genomic DNA, 0.32 μM each primer, 1× final concentration of reaction buffer (20 mM Tris-HCl, 1 mM dithiothreitol [DTT], 0.1 mM EDTA, 0.1 M KCl, 0.5% [vol/vol] Nonidet P40, 0.5% [vol/vol] Tween 20; pH 8.0), 2.5 mM MgCl₂, 200 μM dNTP mixture, and 1 U of Red Diamond Taq polymerase (Eurogentec, Belgium). The thermal cycler (T3 Biometra) was programmed as follows: 5 min at 95°C; 40 cycles of 30 sec at 95°C, 30 sec at 50°C, and 1 min 15 sec at 72°C; and a final extension for 10 min at 72°C. The amplified fragments were sequenced using the two PCR forward primers on an ABI Prism 3100 analyzer (Applied Biosystems) according to the manufacturer’s instructions. The sequences were analyzed using Vector NTI advance software (v11; Invitrogen, Cergy Pontoise, France) to identify the mutations recently shown to be involved in the ex vivo susceptibility to PPQ.

The Mann-Whitney U test was used to compare susceptible and resistant groups. Differences were considered statistically significant if the P value was <0.05. Analyses and graphics were performed using XLSTAT software v19.4 (Addinsoft).

Biobanking of human clinical samples used for malaria diagnostics and secondary uses for scientific purposes is possible as long as the corresponding patients are informed and have not indicated any objections. This requirement was fulfilled here by giving verbal information to the patients, and no immediate or delayed patient opposition was reported to the hospital clinicians. Verbal consent was obtained from all the patients or their parents/guardians before blood collection in Dakar. The ethical committee of the Hôpital Principal de Dakar approved the study.

None of the 76 P. falciparum Senegalese isolates exhibited a bimodal dose-response curve when exposed to PPQ, unlike the Cambodian parasites, and displayed usual sigmoidal dose-response curves (24–26). The distribution of the EC₅₀ values for PPQ and the other six antimalarial drugs is shown in Fig. 1. The ex vivo PPQ EC₅₀ of the 76 isolates ranged from 2.5 to 242 nM. This wide range of PPQ responses is in accordance with previous studies on African parasites assessed using a 42-h isotopic test or a 72-h HRP2 ELISA detection (23, 27, 28). Only the isolates collected in 2016 in Uganda and assessed using a 72-h fluorescence assay with SYBR green I detection presented a narrow range for PPQ EC₅₀, from 1.8 to 26.6 nM (29), unlike those collected in 2010 to 2013, but they were evaluated using HRP2 ELISA detection (3.1 to 188.9 nM) (28). EC₅₀ values depend on methodology, i.e., incubation conditions (gas and time) or detection method (30–33). A limitation could be the use of the standard ex vivo assay to assess PPQ susceptibility and the not PPQ survival assay (PSA) (24). However, PPQ susceptibility is currently more frequently estimated by a standard assay then PSA (15, 25, 28, 29, 34, 35). Amato et al. showed that P. falciparum parasites that were resistant in vitro to PPQ using PSA and associated with amplification of Pfpm2 also had high EC₅₀ values ranging from 89.3 to 159.6 nM using the in vitro standard assay (15).

FIG 1.

Distribution of ex vivo responses of 76 P. falciparum clinical field isolates from Dakar, Senegal, to piperaquine (PPQ), monodesethylamodiaquine (DQ), mefloquine (MQ), chloroquine (CQ), quinine (QN), lumefantrine (LMF), and dihydroartemisinin (DHA). The black line represents the cutoff for ex vivo-reduced susceptibility.

A previous Bayesian analysis on African isolates estimated that parasites with EC₅₀ values greater than 135 nM were considered parasites with reduced susceptibility to PPQ (27). In the present study, 6.6% of the Senegalese isolates were considered parasites with reduced susceptibility to PPQ and showed EC₅₀ values ranging from 139 to 242 nM. In another study on Cambodian isolates, P. falciparum strains resistant ex vivo to PPQ using PSA and associated with amplification of Pfpm2 were also evaluated using the in vitro standard assay (15). These Cambodian P. falciparum strains, which were resistant to PPQ, showed EC₅₀ values ranging from 89.3 to 159.6 nM using the in vitro standard assay, suggesting that parasites with EC₅₀ values above 90 nM were resistant ex vivo to PPQ and associated with in vitro resistance using PSA and amplification of Pfpm2. We deduced a new cutoff at 90 nM for PPQ reduced susceptibility. In the present study, a total of 17.1% of Senegalese isolates showed an EC₅₀ above 90 nM. We analyzed the data from Dakar by considering the two potential thresholds for PPQ reduced susceptibility. These data would suggest that 6.6% to 17.1% of the P. falciparum isolates from Dakar would have reduced susceptibility to PPQ without any current use of DHA-PPQ in Senegal. Significant associations were found between PPQ and DHA (Pearson’s coefficient of correlation of 0.56, P = 0.00002) and MQ (r = 0.46, P = 0.00003). Associations between PPQ and LMF (r = 0.03, P = 0.770), DQ (r = 0.07, P = 0.568), QN (r = 0.05, P = 0.678), or CQ (r = 0.09, P = 0.397) were weak and not significant. Thirty-one percent and 22% of the variation in the ex vivo response to PPQ could be explained by variation in response to DHA and MQ, respectively. A significant association between PPQ and MQ was already estimated with African parasites (r = 0.45) (23). However, the values of the coefficient of determination remain low to fear cross-resistance between PPQ and DHA or MQ and suggest that parasites that have reduced susceptibility to artemisinin derivatives would also be less susceptible to PPQ. Moreover, all the patients infected with parasites with EC₅₀ values ranging between 90 and 100 nM were successfully treated with artesunate or artemether (36). All the patients infected with parasites with EC₅₀ values above 100 nM were treated with quinine. The reduced susceptibility to PPQ observed in this study is not linked to cross-resistance to other quinolone-related drugs, such as amodiaquine or lumefantrine, which are both included in the ACTs that have been used in Senegal for a long time. The values of the coefficient of determination between PPQ and LMF (r² = 0.0009) or PPQ and DQ (r² = 0.0049) were too low to explain the cross-resistance.

The copy number variation of Pfpm2 was successfully analyzed for 64 isolates (11 from 2013, 29 from 2014, and 24 from 2015). The calculated copy numbers ranged from 0.74 to 1.31, with a mean of 0.94 ± 0.09 in the PPQ susceptible group (EC₅₀ under 135 nM) and 0.96 ± 0.05 in the group with PPQ reduced susceptibility (EC₅₀ above 135 nM). Copy numbers were not significantly different between the susceptible and the resistant group (P = 0.286) (Fig. 2). Additionally, there was no difference in copy number variation between P. falciparum isolates with an EC₅₀ under or above 90 nM (0.94 ± 0.11 versus 0.91 ± 0.08).

FIG 2.

Distribution of the estimated copy numbers of Pfpm2 according to piperaquine (PPQ) EC₅₀.

Sequencing of exon 1 and 2 of the exonuclease gene was successfully performed for 76 and 77 isolates, respectively. None of the 77 Senegalese isolates successfully sequenced on exon 2 harbored the exo-E415G mutation, although three other mutations were detected in this gene, including S114C (PPQ IC₅₀, 6.4 nM), N419H (PPQ IC₅₀, 41.1 nM), and T435A (PPQ IC₅₀, 65.6 nM). All isolates with a PPQ EC₅₀ above the threshold of 135 nM were wild type. The E415G mutation and, more generally, polymorphisms in the exonuclease gene in Senegalese isolates are not associated with ex vivo reduced susceptibility to PPQ, contrary to what has been previously shown in Cambodian parasites (15).

None of the isolates showed amplification of Pfpm2, suggesting the absence of PPQ resistance in Dakar. These data are in concordance with those observed in Sierra Leone in 2016 (0%) and in Mozambique in 2015 (1.1%) (37, 38). Rasmussen et al. reported the prevalence of multicopy Pfpm2 of more than 10% in Uganda (29).

Our results show a prevalence of 6.6% (EC₅₀ above 135 nM) to 17.1% (EC₅₀ above 90 nM) of P. falciparum isolates with reduced susceptibility to PPQ in Dakar associated with an absence of plasmepsin II amplification.

Only 10% of P. falciparum isolates collected from clinical failures to DHA-PPQ in Mali carried 2 copies of Pfpm2 (16). Additionally, no correlation was observed between the copy number of Pfpm2 and ex vivo susceptibility to PPQ in P. falciparum Ugandan isolates (29). The three patients harboring parasites with two copies of Pfpm2 were successfully treated with DHA-PPQ in Tanzania (39). The use of DHA-PPQ as an intermittent preventive treatment during pregnancy (IPTp) did not select for genotypes associated with resistance in Cambodia, such as the Pfpm2 or the exonuclease gene (40). No amplification of Pfpm2 and no exo-E415G mutation were found in Cameroonian recrudescent P. falciparum parasites 2 years after treatment by DHA-PPQ (41). Additionally, parasites collected from a malaria case imported from Ethiopia after DHA-PPQ failure had a single copy of Pfpm2 (9). All these data suggest that Pfpm2 is not the only gene involved in PPQ resistance in Africa. Other genes must be studied, such as the RING E3 ubiquitin protein ligase gene (42) or the P. falciparum multidrug resistance 6 gene (Pfmdr6) (43). The K76T mutation in the P. falciparum chloroquine resistance transporter gene (Pfcrt) which is involved in CQ resistance is not involved in in vivo or ex vivo PPQ resistance (44, 45), but new mutations in this gene, such as H97Y, F145I, M343L, or G353V, could also be involved (25, 35).