Role of Pfmdr1 in In Vitro Plasmodium falciparum Susceptibility to Chloroquine, Quinine, Monodesethylamodiaquine, Mefloquine, Lumefantrine, and Dihydroartemisinin

Abstract

The involvement of Pfmdr1 (Plasmodium falciparum multidrug resistance 1) polymorphisms in antimalarial drug resistance is still debated. Here, we evaluate the association between polymorphisms in Pfmdr1 (N86Y, Y184F, S1034C, N1042D, and D1246Y) and Pfcrt (K76T) and in vitro responses to chloroquine (CQ), mefloquine (MQ), lumefantrine (LMF), quinine (QN), monodesethylamodiaquine (MDAQ), and dihydroartemisinin (DHA) in 174 Plasmodium falciparum isolates from Dakar, Senegal. The Pfmdr1 86Y mutation was identified in 14.9% of the samples, and the 184F mutation was identified in 71.8% of the isolates. No 1034C, 1042N, or 1246Y mutations were detected. The Pfmdr1 86Y mutation was significantly associated with increased susceptibility to MDAQ (P = 0.0023), LMF (P = 0.0001), DHA (P = 0.0387), and MQ (P = 0.00002). The N86Y mutation was not associated with CQ (P = 0.214) or QN (P = 0.287) responses. The Pfmdr1 184F mutation was not associated with various susceptibility responses to the 6 antimalarial drugs (P = 0.168 for CQ, 0.778 for MDAQ, 0.324 for LMF, 0.961 for DHA, 0.084 for QN, and 0.298 for MQ). The Pfmdr1 86Y-Y184 haplotype was significantly associated with increased susceptibility to MDAQ (P = 0.0136), LMF (P = 0.0019), and MQ (P = 0.0001). The additional Pfmdr1 86Y mutation increased significantly the in vitro susceptibility to MDAQ (P < 0.0001), LMF (P < 0.0001), MQ (P < 0.0001), and QN (P = 0.0026) in wild-type Pfcrt K76 parasites. The additional Pfmdr1 86Y mutation significantly increased the in vitro susceptibility to CQ (P = 0.0179) in Pfcrt 76T CQ-resistant parasites.

INTRODUCTION

Over the past 20 years, many strains of Plasmodium falciparum have become resistant to chloroquine (CQ) and other antimalarial drugs. In response to increasing CQ resistance, Senegal switched, in 2004, to sulfadoxine-pyrimethamine with amodiaquine as the first-line therapy. In 2006, the Senegalese National Malaria Control Programme recommended artemisinin-based combination therapy (ACT) as the first-line treatment for uncomplicated malaria. The combination sulfadoxine-pyrimethamine and amodiaquine treatment was changed to artemether-lumefantrine and artesunate-amodiaquine. Since 2006, more than 1.5 million ACT-based treatments have been administered in Senegal (1). In 2006, the Senegalese National Malaria Control Programme also recommended testing for all suspected cases of malaria with the P. falciparum histidine-rich protein 2 (PfHRP2)-based rapid diagnostic test (RDT). Since this time, ACT use has been restricted to confirmed malaria cases to reduce drug pressure. In 2009, 184,170 doses of ACT were dispensed in Senegal (2).

The ability to maximize the efficacy and longevity of antimalarial drugs for malaria control will depend critically on intensive research to identify in vitro markers along with ex vivo and in vivo surveillance programs. Furthermore, it is necessary to identify molecular markers that predict antimalarial resistance or decreased susceptibility so that active surveillance can monitor temporal trends in parasite susceptibility. Although the molecular mechanisms underlying multidrug resistance by P. falciparum remain largely unknown, polymorphisms within the Pfmdr1 (Plasmodium falciparum multidrug resistance 1) gene, which encodes a transmembrane homolog of the PGH1 protein, have been implicated. Field work has shown that the predictive value of CQ resistance and point mutations in the Pfmdr1 sequence resulting in amino acid changes varies depending on the geographic area (3, 4). Five point mutations have been described: N86Y, Y184F, S1034C, N1042D, and D1246Y. Point mutations, most notably 86Y, have been associated with a decrease in CQ susceptibility (5). However, in some epidemiological studies, the number of CQ-susceptible samples is too limited to provide a statistically meaningful analysis (4, 6). Using precautions, no relationship or only weak relationships have been established between CQ resistance and mutations in Pfmdr1 in P. falciparum (3). However, the risk of therapeutic failure with CQ is greater for patients harboring the 86Y mutation, with an odds ratio (OR) of 2.2 (95% confidence interval [CI], 1.6 to 3.1) for a 14-day follow-up and 1.8 (95% CI, 1.3 to 2.4) for a 28-day follow-up (7).

In addition, the risk of therapeutic failure with amodiaquine is greater for patients harboring the 86Y mutation, with an OR of 5.4 (95% CI, 2.6 to 11.2; meta-analysis of six studies) (7); this mutation increases the risk of failure with amodiaquine plus sulfadoxine-pyrimethamine with an OR of 7.9 (8).

It has been shown through heterologous expression that Pfmdr1 mutations at codons 1034 and 1042 abolish or reduce the level of resistance to mefloquine (MQ) (9). Moreover, transfection with a wild-type Pfmdr1 allele at codons 1034, 1042, and 1246 confers MQ resistance to susceptible parasites (10). However, mutations at Pfmdr1 codons 1034, 1042, and 1246 in P. falciparum isolates are not sufficient to explain the variations in MQ susceptibility (11). Analyses of P. falciparum isolates show an association between mutation at codon 86 and an increase in susceptibility to MQ, halofantrine, or artemisinin derivatives (12–14). The selection of Pfmdr1 polymorphisms by the combination of artemether and lumefantrine recently has been observed (15). Mutation in Pfmdr1 also has been associated with decreased susceptibility to artemether and lumefantrine (LMF) drugs separately (16–18).

However, mutations at the Pfmdr1 gene do not seem to be sufficient to explain in vitro resistance to antimalarial drugs, and additional gene mutations are necessary.

The Pfcrt gene was first identified in 2000 (19). To date, at least 20 point mutations have been described (19–21), but only one is the reference mutation (K76T), which is a marker of the CQ-resistant phenotype. This mutation often is associated with other mutations in the Pfcrt gene, whose role is not yet defined. The odds ratio for CQ failure associated with the K76T mutation was 2.1 (95% confidence interval, 1.5 to 3.0; meta-analysis of 13 studies) for a 14-day follow-up and 7.2 (95% CI, 4.5 to 11.5; meta-analysis of 12 studies) for a 28-day follow-up (7). However, the existence of CQ-susceptible strains associated with the K76T mutation suggests that other genes could be involved in the resistance to CQ. The 76T mutation is necessary, but not sufficient, for influencing CQ susceptibility (22).

The aim of this study was first to evaluate the association between polymorphisms in Pfmdr1 and in vitro responses to CQ, MQ, LMF, quinine (QN), monodesethylamodiaquine (the metabolite of amodiaquine) (MDAQ), and dihydroartemisinin (DHA) in Senegal. The association between Pfcrt K76T and Pfmdr1 mutations and in vitro susceptibility to the six drugs then was investigated.

MATERIALS AND METHODS

Patients and sample collection.

P. falciparum isolates from patients with malaria who live in Dakar (>80%) and the surrounding area and did not travel during the previous month were obtained during the rainy seasons of October 2009 to January 2010 (172 patients, 42% female) and August 2010 to January 2011 (129 patients, 38% female). The patients with malaria were recruited at the Hôpital Principal de Dakar, a military hospital, within the context of an evaluation of ex vivo malaria susceptibility to antimalarial drugs in Dakar (23, 24). Venous blood samples were collected in Vacutainer ACD tubes (Becton Dickinson, Rutherford, NJ, USA) prior to patient treatment. Of the 301 patients, 54% were recruited from the emergency department during each of the 2 seasons; other patients were recruited from the intensive care unit (18% during October 2009 to January 2010 and 20% during August 2010 to January 2011), pediatric department (9% and 5%), and other units (19% and 21%). No significant differences were found between the 2 seasons for parasitemia (P = 0.160), sex ratio (P = 0.446), area of residence (P = 0.651), or hospital admission status (P = 0.567). Information on antimalarial treatment prior to admission was not available. Informed verbal consent from the patients and/or their parents/guardians was obtained before blood collection; the study was approved by the ethical committee of the Hôpital Principal de Dakar. An assessment of P. falciparum susceptibility to antimalarial drugs was performed using the same venous blood sample as that used for this diagnostic analysis.

Thin blood smears were stained using a RAL kit (Réactifs RAL, Paris, France) and examined to determine the P. falciparum density and confirm monoinfection. Parasitized erythrocytes were washed three times with RPMI 1640 medium (Invitrogen, Paisley, United Kingdom) buffered with 25 mM HEPES and 25 mM NaHCO₃. If parasitemia exceeded 0.5%, the infected erythrocytes were diluted to 0.5% with uninfected erythrocytes (human blood type A⁺) and resuspended in RPMI 1640 medium supplemented with 10% human serum (Abcys S.A., Paris, France) for a final hematocrit of 1.5%.

Drugs.

CQ, QN, and DHA were purchased from Sigma (St. Louis, MO, USA). MDAQ was obtained from the World Health Organization (Geneva, Switzerland), MQ was purchased from Roche (Paris, France), and LMF was purchased from Novartis Pharma (Basel, Switzerland). QN, MDAQ, MQ, and DHA first were dissolved in methanol and then diluted in water to final concentrations ranging from 5 nM to 3,200 nM for QN, 1.56 nM to 1,000 nM for MDAQ, 3.2 nM to 400 nM for MQ, and 0.1 nM to 100 nM for DHA. CQ was resuspended and diluted in water to final concentrations ranging from 5 nM to 3,200 nM. LMF was resuspended and diluted in ethanol to obtain final concentrations ranging from 0.5 nM to 310 nM.

Batches of plates were tested and validated using the chloroquine-susceptible strain 3D7 (isolated in West Africa; obtained from MR4, VA, USA) and the chloroquine-resistant W2 (isolated in Indochina; obtained from MR4, VA, USA) in three to six independent experiments using the conditions described below. The two strains were synchronized twice with sorbitol before use (25), and clonality was verified every 15 days through PCR genotyping of the polymorphic genetic markers msp1 and msp2 and microsatellite loci (26, 27). Additionally, clonality was verified each year by an independent laboratory from the Worldwide Antimalarial Resistance Network (WWARN).

Ex vivo assay.

For in vitro isotopic microtests, 200 μl of parasitized red blood cells (final parasitemia, 0.5%; final hematocrit, 1.5%) was aliquoted into 96-well plates predosed with antimalarial drugs. The plates were incubated in a sealed bag for 42 h at 37°C with atmospheric generators for capnophilic bacteria (Genbag CO₂) at 5% CO₂ and 15% O₂ (bioMérieux; Marcy l'Etoile, France) (28). After thawing the plates, the hemolysed cultures were homogenized by vortexing the plates. Both the success of the drug susceptibility assay and the appropriate volume of hemolysed culture to use for each assay were determined for each clinical isolate during a preliminary pLDH enzyme-linked immunosorbent assay (ELISA). Both the pretest and subsequent experimental ELISAs were performed using a commercial kit (ELISA-Malaria antigen test, reference 750101; DiaMed AG, Cressier s/Morat, Switzerland), as previously described (29). The optical density (OD) of each sample was measured with a spectrophotometer (Multiskan EX; Thermo Scientific, Vantaa, Finland).

The concentration at which the drugs were able to inhibit 50% of parasite growth (IC₅₀) was calculated using the inhibitory sigmoid maximum effect (E_max) model with an estimation of the IC₅₀ through nonlinear regression using a standard function of the R software (ICEstimator, version 1.2) (30). The IC₅₀s were validated only if the OD ratio (OD at concentration 0/OD at maximum concentration) was higher than 1.8 and the confidence interval ratio (upper 95% confidence interval of the IC₅₀ estimation/lower 95% confidence interval of the IC₅₀ estimation) was lower than 2.0 (30).

Nucleic acid extraction.

The total genomic DNA of each strain was isolated using the QIAampW DNA minikit according to the manufacturer's recommendations (Qiagen, Germany).

Pfmdr1 single-nucleotide polymorphisms.

Two primer pairs were used to amplify Pfmdr1 fragments carrying the five key codons (31). A 590-bp fragment was amplified with a primer pair (sense, 5′-AGA GAA AAA AGA TGG TAA CCT CAG-3′; antisense, 5′-ACC ACA AAC ATA AAT TAA CGG-3′) to determine the sequences of codons 86 and 184 (MDR1-1), and a second fragment (968 bp) was amplified with a primer pair (sense, 5′-CAG GAA GCA TTTTAT AAT ATG CAT-3′; antisense, 5′-CGT TTAACA TCT TCC AAT GTT GCA-3′) to determine the sequences of codons 1034, 1042, and 1246 (MDR1-2) (31). The reaction mixture consisted of approximately 2.5 μl of genomic DNA, 0.5 μM forward and reverse primers, 2.5 μl of 10× reaction buffer (Eurogentec), 2.5 mM MgCl₂, 200 μM deoxynucleoside triphosphate mixture (dGTP, dATP, dTTP, and dCTP) (Euromedex, Souffelweyersheim, France), and 1 U of Red Gold StarW DNA polymerase (Eurogentec) in a final volume of 25 μl. The thermal cycler (T3 Biometra) was programmed for MDR1-1 with an initial step at 94°C for 5 min; 40 cycles of 94°C for 30 s, 52°C for 30 s, and 72°C for 1 min; and a final 10-min extension step at 72°C. For MDR1-2, the parameters were an initial step at 94°C for 5 min; 40 cycles of 94°C for 30 s, 56°C for 1 min, and 72°C for 90 s; and a final 10-min extension step at 72°C. The PCR products were separated using a 1.5% agarose gel containing 0.5 μg/ml ethidium bromide. Amplicons were purified using the QIAquick96 PCR BioRobot kit and an automated protocol on the BioRobot 8000 workstation (Qiagen, Courtaboeuf, France). The purified fragments were sequenced using the BigDye Terminator v3.1 cycle sequencing kit (Applied Biosystems) using the primers described above. The sequencing reaction products were purified using the BigDye XTerminatorW purification kit (Applied Biosystems) in accordance with the manufacturer's instructions. The purified products were sequenced using an ABI Prism 3100 analyzer (Applied Biosystems), and the sequences were analyzed using VectorNTI advance software (version 11; Invitrogen, Cergy Pontoise, France).

Pfcrt single-nucleotide polymorphisms.

A 546-nucleotide fragment of the Pfcrt gene (containing codon 76) was amplified by PCR using CRTP1-sense (5′-CCG TTA ATA ATA AAT ACA CGC AG-3′) and CRTP1-antisense (5′-CGG ATG TTA CAA AAC TAT AGT TAC C-3′) primers (32). The reaction mixture for PCR amplifications included 2.5 μl of genomic DNA, 2.5 μl of 10× reaction buffer (Eurogentec), 0.5 μM each primer, 200 μM a deoxynucleoside triphosphate mixture (dGTP, dATP, dTTP, and dCTP) (Euromedex, Souffelweyersheim, France), 2.5 mM MgCl₂, and 1 U of RedGoldStar DNA polymerase (Eurogentec) in a final volume of 25 μl. The thermal cycler (T3 Biometra, Archamps, France) was programmed with an initial 94°C incubation for 5 min; 40 cycles of 94°C for 20 s, 56°C for 20 s, and 60°C for 40 s; and a final 5-min extension step at 60°C. The PCR products were loaded on a 1.5% agarose gel containing 0.5 μg/ml ethidium bromide. The PCR products were diluted 1:100 in distilled water, and 2.5 μl of the final dilution was used for the second PCR. This PCR amplified a 275-bp segment around the mutation using a common inner primer, CRTP3-sense (5′-TGA CGA GCG TTA TAG AG-3′), coupled with either CRTP4m-antisense (5′-GTT CTT TTA GCA AAA ATT G-3′) (detects the 76T codon) or CRTP4w-antisense (5′-GTT CTT TTA GCA AAA ATT T-3′) (detects the 76K codon) (20). The reaction mixture for the PCR amplifications included 2.5 μl of diluted PCR product, 2.5 μl of 10× reaction buffer (Eurogentec), 0.5 μM each primer, 200 μM deoxynucleoside triphosphate mixture, 1.5 mM MgCl₂, and 0.75 U of RedGoldStar DNA polymerase (Eurogentec) in a final volume of 25 μl. The PCR conditions were an initiation at 94°C for 5 min; 15 cycles at 94°C for 20 s, 48.5°C for 20 s, and 64°C for 40 s; and a final 5-min extension step at 64°C. Purified genomic DNA from P. falciparum clones 3D7 (CQ susceptible) and W2 (CQ resistant) were used as positive controls, and water and human DNA were used as negative controls. The PCR products from the amplification reactions were evaluated by electrophoresis on 2% agarose gels.

Statistical analysis.

The data were analyzed using R software (version 2.10.1). Differences between the IC₅₀s of the isolates and the Pfmdr1 and/or Pfcrt polymorphisms were compared using the Wilcoxon rank sum test, the Kruskal-Wallis test, and the Welch two-sample t test.

RESULTS

Of the 301 isolates tested for Pfmdr1 polymorphisms, 174 were successfully evaluated by ex vivo testing. The average parameter estimates for the 6 antimalarial drugs used against the P. falciparum isolates are given in Table 1.

TABLE 1.

Ex vivo susceptibility of 174 Plasmodium falciparum isolates from Dakar to the studied drugs

Drug	No. of isolates	Mean IC50 (nM)	95% CI	IC50 range
				Min	Max
CQ	174	97.7	52.3–143.0	3.7	1,958
MDAQ	174	28.3	23.3–33.3	0.85	274.7
LMF	174	22.9	18.8–26.9	0.45	149.6
QN	173	283.9	231.4–336.4	4.8	1,291
MQ	173	38.3	33.8–42.9	2.4	170.3
DHA	163	2.9	2.5–3.3	0.1	14.0

The Pfmdr1 86Y mutation was identified in 14.9% of the samples (26 out of 174 isolates), and the 184F mutation was identified in 71.8% of the isolates (125 out of 174). No 1034C, 1042N, or 1246Y mutations were detected.

The Pfmdr1 86Y mutation was significantly associated with increased susceptibility to MDAQ (P = 0.0023 by Wilcoxon rank sum test), LMF (P = 0.0001), DHA (P = 0.0387), and MQ (P = 0.00002). For the 86Y mutation isolates, the IC₅₀s were significantly reduced for MDAQ (14.4 nM versus 30.5 nM; P = 0.00021), LMF (9.2 nM versus 25.0 nM; P = 0.00019), MQ (18.6 nM versus 42.0 nM; P < 0.00001), and DHA (2.1 nM versus 3.0; P = 0.049) (Table 2). The N86Y mutation was not associated with CQ (P = 0.214) or QN (P = 0.287) responses.

TABLE 2.

Ex vivo susceptibility of 174 Plasmodium falciparum isolates from Dakar to the studied drugs according to isolates with the 86Y pfmdr1 and 184Y pfmdr1 mutations

Drug	Mean IC50 (nM) for isolates with:											P valuea for 86Y/86Y-Y184 haplotype
	Pfmdr1 N86 (n = 141 to 148)c	Pfmdr1 86Y (n = 22 to 26)	P valuea	Pfmdr1 Y184 (n = 45 to 49)	Pfmdr1 184F (n = 119 to 125)	P valuea	Pfmdr1 N86, Pfmdr1 Y184 (n = 45 to 48)	Pfmdr1 N86, Pfmdr1 184F (n = 96 to 101)	Pfmdr1 86Y, Pfmdr1 Y184 (n = 2)	Pfmdr1 86Y, Pfmdr1 184F (n = 22 to 24)	P valueb
CQ	105.0	49.9	0.0572	92.3	99.9	0.776	95.7	109.5	12.5	53.7	0.0579	0.0137
MDAQ	30.5	14.4	0.0002	32.1	26.8	0.466	33.4	29.2	3.8	15.4	0.0136	0.0121
LMF	25.0	9.2	0.0002	26.7	21.3	0.281	27.6	23.7	5.7	9.5	0.0019	0.3210
QN	295.2	211.6	0.1150	219.6	310.7	0.0572	220.2	331.5	206.1	212.2	0.1201	0.9100
MQ	42.0	18.6	<0.00001	34.0	40.0	0.220	35.1	45.2	9.0	19.4	0.0001	0.3191
DHA	3.0	2.1	0.0495	2.7	2.9	0.762	2.8	3.1	2.3	2.1	0.0970	0.9811

^aDetermined by Welch's two-sample t test.

^bDetermined by Kruskal-Wallis test.

^cn = number of tested isolates for each drug.

The Pfmdr1 184F mutation was not associated with variable susceptibility responses to the 6 antimalarial drugs (P = 0.168 for CQ, 0.778 for MDAQ, 0.324 for LMF, 0.961 for DHA, 0.084 for QN, and 0.298 for MQ; all by Wilcoxon rank sum test). There was no significant difference in IC₅₀ between the isolates with Y184 or 184F (P values from 0.0572 to 0.776; Welch's two-sample t test) (Table 2).

The IC₅₀ distribution for the four haplotypes (N86-Y184, N86-F184, 86Y-Y184, and 86Y-184F) is presented in Fig. 1. The 86Y-Y184 haplotype was significantly associated with increased susceptibility to MDAQ (mean IC₅₀, 3.8 nM; P = 0.0136; Kruskal-Wallis test), LMF (mean IC₅₀, 5.7 nM; P = 0.0019), and MQ (mean IC₅₀, 9.0 nM, P = 0.0001) (Table 2).

FIG 1.

Relationship between chloroquine (CQ), quinine (QN), monodesethylamodiaquine (MDAQ), lumefantrine (LMF), mefloquine (MQ), and dihydroartemisinin (DHA) and the Pfmdr1 86 and 184 haplotypes.

Only CQ responses were significantly associated with Pfcrt 76T. The CQ IC₅₀s were significantly increased in the group of isolates with the 76T mutation (53.4 nM versus 166.7 nM; P = 0.0002 by Welch's two-sample t test) (Table 3).

TABLE 3.

Ex vivo susceptibility of 174 Plasmodium falciparum isolates from Dakar to the studied drugs according to isolates with the 86Y Pfmdr1 and 86T Pfcrt mutations

Drug	Mean IC50 (nM) for isolates with:											P valuea for K76/86Y-K76, 86Y/86Y-K76 haplotype
	Pfmdr1 N86 (n = 141 to 148)c	Pfmdr1 86Y (n = 22 to 26)	P valuea	Pfcrt K76 (n = 97 to 101)	Pfcrt 76T (n = 64 to 70)	P valuea	Pfmdr1 N86, Pfcrt K76 (n = 81 to 86)	Pfmdr1 N86, Pfcrt 76T (n = 54 to 59)	Pfmdr1 86Y, Pfcrt K76 (n = 14 to 16)	Pfmdr1 86Y, Pfcrt 76T (n = 8 to 10)	P valueb
CQ	105.0	49.9	0.0572	53.4	166.7	0.0002	56.4	179.7	35.1	75.9	0.0301	0.2247
MDAQ	30.5	14.4	0.0002	27.8	29.1	0.1603	31.2	29.4	7.1	27.2	0.00002	<0.00001, 0.0458
LMF	25.0	9.2	0.0002	24.7	19.9	0.2219	27.7	20.6	6.0	14.8	0.00001	<0.00001, 0.3584
QN	295.2	211.6	0.1150	277.1	295.0	0.1107	300.9	285.9	129.3	355.7	0.0172	0.0026, 0.1160
MQ	42.0	18.6	<0.00001	40.5	35.1	0.3631	46.2	36.2	12.3	28.6	0.00002	<0.00001, 0.1074
DHA	3.0	2.1	0.0495	2.4	3.6	0.0710	2.5	3.7	2.0	2.7	0.0585	0.3332

^aDetermined by Welch's two-sample t test.

^bDetermined by Kruskal-Wallis test.

^cn = number of tested isolates for each drug.

The IC₅₀s for the Pfcrt K76-Pfmdr1 86Y haplotype combination were significantly lower for MDAQ (P = 0.00002 by Kruskal-Wallis test), LMF (P = 0.00001), MQ (P = 0.00002), and QN (P = 0.0172) (Table 3). The CQ IC₅₀s were significantly higher for the Pfcrt 76T-Pfmdr1 N86 haplotype combination (P = 0.0301) (Table 3).

The additional Pfmdr1 86Y mutation significantly increased the in vitro susceptibility to MDAQ (P < 0.0001), LMF (P < 0.0001), MQ (P < 0.0001), and QN (P = 0.0026) in wild-type Pfcrt K76 parasites (Table 3). The additional Pfmdr1 86Y mutation significantly increased the in vitro susceptibility to CQ (P = 0.0179) in Pfcrt 76T CQ-resistant parasites.

The association between Pfmdr1 Y184F and Pfcrt K76T did not significantly influence the IC₅₀ for CQ, MDAQ, LMF, QN, MQ, or DHA (Table 4).

TABLE 4.

Ex vivo susceptibility of 174 Plasmodium falciparum isolates from Dakar to the studied drugs according to isolates with the 184F Pfmdr1 and 86T Pfcrt mutations

Drug	Mean IC50 (nM) for isolates with:
	Pfmdr1 Y184 (n = 45 to 49)c	Pfmdr1 184F (n = 119 to 125)	P valuea	Pfcrt K76 (n = 97 to 101)	Pfcrt 76T (n = 64 to 70)	P valuea	Pfmdr1 Y184, Pfcrt K76 (n = 31 to 33)	Pfmdr1 Y184, Pfcrt 76T (n = 14 to 15)	Pfmdr1 184F, Pfcrt K76 (n = 60 to 68)	Pfmdr1 184F, Pfcrt 76T (n = 47 to 52)	P valueb
CQ	92.3	99.9	0.776	53.4	166.7	0.0002	44.5	197.5	57.9	157.3	0.0013
MDAQ	32.1	26.8	0.466	27.8	29.1	0.1603	33.0	30.2	25.4	28.8	0.4001
LMF	26.7	21.3	0.281	24.7	19.9	0.2219	27.1	25.7	23.5	18.1	0.7365
QN	219.6	310.7	0.057	277.1	295.0	0.1107	200.4	253.1	312.4	308.3	0.1524
MQ	34.0	40.0	0.220	40.5	35.1	0.3631	32.1	38.1	44.5	34.3	0.3925
DHA	2.7	2.9	0.762	2.4	3.6	0.0710	2.6	3.3	2.2	3.6	0.7365

^aDetermined by Welch's two-sample t test.

^bDetermined by Kruskal-Wallis test.

^cn = number of tested isolates for each drug.

DISCUSSION

Maximizing the efficacy and longevity of antimalarial drugs remains important and will critically depend on the pursuit of intensive research toward the identification of in vitro markers of resistance and the implementation of ex vivo and in vivo surveillance programs, such as those championed by the Worldwide Antimalarial Resistance Network (33, 34). Specifically, there is a need to identify molecular markers that effectively predict antimalarial drug resistance and enable the active surveillance of temporal trends in parasite susceptibility (35).

The role of polymorphism in Pfmdr1 is still debated. However, single-nucleotide polymorphisms in Pfmdr1 have been shown to be associated with differential in vivo efficacy and in vitro P. falciparum susceptibility to ACT antimalarial partner drugs, including amodiaquine (36, 37), mefloquine (10, 38), lumefantrine (39, 40), and artemisinin derivatives (13, 14, 38).

Pfmdr1 mutations at codons 86, 1034, 1042, and 1246 have been associated with in vitro resistance to MQ and artesunate in genetically modified parasite lines (10, 38) and clinical isolates (18, 41). It has been shown through heterologous expression that Pfmdr1 mutations at codons 1034 and 1042 abolish or reduce the level of resistance to MQ (9). Moreover, transfection with a wild-type Pfmdr1 allele at codons 1034, 1042, and 1246 confers MQ resistance to susceptible parasites (10). However, mutations at codons 1034, 1042, and 1246 in P. falciparum Pfmdr1 isolates are not sufficient to explain variations in MQ susceptibility (11). Analyses of P. falciparum isolates revealed an association between mutation at codon 86 and an increase in susceptibility to MQ (12–14). However, as no mutation at codons 1034, 1042, and 1246 was found in the 174 P. falciparum isolates from Dakar, these polymorphisms cannot explain the variation in in vitro MQ susceptibility in Senegal. Nonetheless, we show that the 86Y mutation was significantly associated with increased susceptibility to MQ (P = 0.00002). These results are in accordance with previous studies in Asia (12, 42) and in Benin (43) and are strengthened by field studies demonstrating MQ selection of the N86 allele in recurrent infections after treatment with artesunate plus MQ (44).

We show that the N86 allele can predict in vitro decreased susceptibility to LMF, whereas the 86Y mutation was significantly associated with increased susceptibility to LMF (P = 0.0001). These results are in accordance with previous in vitro studies in Asia (45), Kenya (17), and Benin (43). Field studies in east Africa also have shown selection of the 86N allele in recurrent infections after treatment with artemether plus LMF (39, 40, 46, 47), which suggests that 86N is a marker of LMF resistance in vivo.

We show that the Pfmdr1 N86 allele can predict in vitro decreased susceptibility to MDAQ, the metabolite of amodiaquine; in contrast, the 86Y mutation was significantly associated with increased susceptibility to MDAQ (P = 0.0023). However, these data are in contrast to previous in vitro works. There was no difference in MDAQ IC₅₀ between the two haplotypes N86 and 86Y in isolates from Benin (43). In other works, the mutant Pfmdr1 86Y allele showed an increased MDAQ IC₅₀ in isolates from Nigeria (48). The Pfmdr1 86Y mutation has been shown to be associated with treatment failure after monotherapy with amodiaquine (36, 49) or after combination therapy with artesunate-amodiaquine (50). In a meta-analysis, the Pfmdr1 86Y mutation was found to be associated with amodiaquine failure, with an odds ratio of 5.4 (7). The Pfmdr1 1246Y mutation also has been found to be associated with in vitro resistance to amodiaquine (51) and with recrudescent infection after treatment with amodiaquine or amodiaquine-artesunate (49, 50).

Here, we show that the N86 allele can predict in vitro decreased susceptibility to DHA and that the 86Y mutation was significantly associated with increased susceptibility to DHA (P = 0.0387). These results are in accordance with previous in vitro studies on parasite laboratory strains (13, 14, 38) or in clinical isolates (44). The Pfmdr1 86Y mutation has been shown to be associated with in vivo resistance to ACT after combination therapy with artesunate-amodiaquine (50) or artemether plus LMF (39, 40, 46, 47), which suggests that 86N is a marker of artemisinin derivative resistance in vivo. However, recently it has been shown that this type of in vitro test is not adapted to detect resistance to artemisinin derivatives. A nonstandard phenotypic test currently is used and recommended for evaluating ex vivo or in vitro susceptibility to artemisinin derivatives correlating with the clearance delay of parasites (52, 53). In addition, mutations in a new gene, K13-propeller, are associated with in vitro parasite survival rates, as determined by this new assay and in vivo parasite clearance rates (54).

The N86Y mutation was not associated with QN (P = 0.287). The involvement of N86Y in QN resistance is still debated: in some studies, the mutation 86Y is associated with increased susceptibility (42, 55), although this is not the case in other studies (56). The 1042D mutation has been reported to induce QN resistance (38).

The N86Y mutation was not associated with CQ responses (P = 0.214) in Senegalese P. falciparum isolates. These results are in accordance with previous studies in Kenya (17) and Nigeria (57). However, the 86Y mutation has been associated with a decrease in CQ susceptibility in some works (5). However, in some of these epidemiological studies, the number of CQ-susceptible samples is too limited to provide a statistically meaningful analysis (4, 6). Using precautions, no relationship or only weak relationships are established between CQ resistance and mutations in Pfmdr1 in P. falciparum (3).

The 184F mutation was not associated with various susceptibility responses to the 6 antimalarial drugs tested (P = 0.168 for CQ, 0.778 for MDAQ, 0.324 for LMF, 0.961 for DHA, 0.084 for QN, and 0.298 for MQ). Thus, the significance of the 184F mutation remains less well understood. Indeed, no clear association between the 184F mutation and MQ failure has been established. A study showed that Asian isolates with a single 184F mutation exhibited increased resistance to MQ and AS (41). A study from Cambodia demonstrated that isolates with a single 184F mutation had significantly increased IC₅₀s for MQ (58). Similar results were obtained in studies from Tanzania, Uganda, and Nigeria, showing the selection of 86N, 184F, and 1246D Pfmdr1 alleles in recurrent infections after treatment with artemether plus LMF (39, 47, 59, 60). In addition, the 184F mutation was shown to be associated with in vitro resistance to QN (55). In our study, the 184F mutation was not associated with QN susceptibility responses (P = 0.084), although the P value was not very distant from the limit of significance. The mean IC₅₀ was 219.6 nM for the Y184 allele and 310.7 nM for the 184F allele (P = 0.057).

Evaluation of the Pfmdr1 86 plus 184 haplotype showed a significantly increased in vitro susceptibility to MDAQ, LMF, and MQ in parasites with 86Y plus Y184. A similar association was seen for LMF and MQ in previous studies in Benin (43). Similarly, we showed a decreased susceptibility to LMF and MQ in parasites with N86 plus 184F and N86 plus Y184. The parasites with Pfmdr1 86Y plus Y184 and 86Y plus 184F showed a nonsignificant increase of CQ susceptibility (P = 0.0579). However, univariate analysis showed that there was a significant association between CQ treatment failure and the presence of Pfmdr1 86Y plus Y184 in Nigeria (57).

The IC₅₀s for the Pfcrt K76-Pfmdr1 86Y haplotype combination were significantly lower for MDAQ (P = 0.00002), LMF (P = 0.00001), MQ (P = 0.00002), and QN (P = 0.0172). The additional Pfmdr1 86Y mutation significantly increased the in vitro susceptibility to MDAQ (P < 0.0001), LMF (P < 0.0001), MQ (P < 0.0001), and QN (P = 0.0026) in wild-type Pfcrt K76 parasites from Senegal. The LMF IC₅₀ was significantly lower for Pfmdr1 86Y-Pfcrt K76 (mean IC₅₀, 6.0) and Pfmdr1 86Y-Pfcrt 76T (mean IC₅₀, 14.8). These results are in concordance with previous data from Kenya (17). Artemether plus LMF is used largely in Senegal, and Pfmdr1 N86 certainly is selected in recurrent infection after treatment with artemether plus LMF, like in east Africa (34, 35, 46, 47). For MQ, the 86Y-K76 haplotype showed a lower IC₅₀ (mean IC₅₀, 12.3 nM), in concordance with previous work in western Kenya (61). For MDAQ, the 86Y-K76 haplotype showed a lower IC₅₀ (mean IC₅₀, 7.1 nM). In western Kenya, the 86Y-K76 haplotype showed the highest IC₅₀ (61). In southern Sudan, more than 80% of the MDAQ-susceptible isolates carried the wild-type Pfcrt K76-Pfmdr1 N86 genotype (62). The 86Y-76T haplotype was found to be associated with recrudescence infection after treatment with amodiaquine (49, 63). The CQ IC₅₀s were significantly higher for the Pfcrt 76T-Pfmdr1 N86 haplotype combination (P = 0.0301). The N86-K76 and 86Y-K76 haplotypes showed a lower CQ IC₅₀, in concordance with previous work in Kenya (17, 61). The additional Pfmdr1 86Y mutation significantly increased the in vitro susceptibility to CQ (P = 0.0179) in Pfcrt 76T CQ-resistant parasites from Senegal, while the 86Y-76T haplotype was shown to be associated with in vitro CQ resistance in isolates from Nigeria (57).

The Y184-76T and 184F-76T haplotypes showed the highest CQ IC₅₀ but without any significant difference between the two haplotypes in Dakar, in concordance with previous data from Papua New Guinea (64). The association between Pfmdr1 Y184F and Pfcrt K76T did not significantly influence the IC₅₀ for CQ, MDAQ, LMF, QN, MQ, or DHA.

Another molecular marker of antimalarial drug susceptibility is the amplification of Pfmdr1; an increase in the copy number of Pfmdr1 is associated with clinical failure and with in vitro resistance to aryl-amino-alcohols, particularly MQ, QN, and LMF (44, 65). Pfmdr1 amplification also has been demonstrated to decrease the susceptibility to artemisinin derivatives in the field as well as in vitro (65, 66). However, the role of the amplification of Pfmdr1 in P. falciparum resistance to antimalarial drugs in Africa is debated. We previously analyzed the number of Pfmdr1 copies in these 174 P. falciparum samples, and a significant association was found only for dihydroartemisinin (P = 0.003) (67). The odds ratio for reduced in vitro susceptibility to dihydroartemisinin associated with 2 Pfmdr1 copies was 1.4. However, the number of isolates with duplicated Pfmdr1 copies was small (9 out of 174 samples).

In summary, we demonstrated that the 86Y mutation is significantly associated with increased susceptibility to MDAQ, LMF, MQ, and DHA in P. falciparum parasites from Dakar, and the Pfmdr1 86Y-Y184 haplotype significantly increases the in vitro susceptibility to MDAQ, LMF, and MQ in parasites. The additional Pfmdr1 86Y mutation increases significantly the in vitro susceptibility to MDAQ, LMF, MQ, and QN in wild-type Pfcrt K76 parasites and to CQ in Pfcrt 76T CQ-resistant parasites from Senegal.