High-level artemisinin-resistance with quinine co-resistance emerges in P. falciparum malaria under in vivo artesunate pressure

Four- to six-week-old male and female NOD/SCID IL-2Rγ^−/− (NSG) mice (Charles River, France) were housed in sterile isolators and supplied autoclaved tap water with a γ-irradiated pelleted diet ad libitum. They were manipulated under pathogen-free conditions using a laminar-flux hood.

HuRBC were used as host-cells for all in vitro and in vivo experiments. Packed huRBC were provided by the French Blood Bank (Etablissement Français du Sang, France) and taken from donors with no history of Malaria. HuRBC were suspended in SAGM (Saline, Adenine, Glucose, Mannitol solution) and kept at 4 °C for a maximum of 2 weeks. Before injection, huRBC were washed thrice in RPMI-1640 medium (Gibco-BRL, Grand Island, NY, USA) supplemented with 1 mg of hypoxanthine per liter (Sigma-Aldrich, St Louis, MO, USA) and warmed for 10 min to 37 °C.

The P. falciparum Uganda Palo Alto Marburg strain (FUP/CB or PAM) was used for all experiments [33]. This pan-sensitive strain is used as a laboratory reference for antimalarial assays [34, 35]. Over time, strains with different levels of ART-R were cryopreserved using the glycerol/sorbitol method as described [36]. Parasites were cultured in vitro with 5% hematocrit, at 37 °C with 5% CO₂, using RPMI-1640 medium (Gibco-BRL) with 35 mM HEPES (Sigma-Aldrich), 24 mM NaHCO₃, 10% albumax (Gibco-BRL), and 1 mg/L of hypoxanthine (Sigma-Aldrich). When required, cultures were synchronized by either plasmagel (Roger Bellon, Neuilly-sur-Seine, France) flotation [37] or exposure to 5% sorbitol (Sigma-Aldrich) [38]. At regular intervals, cultures were tested for Mycoplasma contamination using PCR.

P. falciparum was maintained in huRBC grafted in NSG immunocompromised mice undergoing additional modulation of innate defenses using clodronate-containing liposomes, as described previously [31, 32] (‘Pf-NSG’ model). The proportion of huRBC in mouse blood (chimerism) was measured during experiments every 6 ± 4.5 days (mean ± standard deviation (SD)) by flow cytometry (Facscalibur, BD Biosciences, Franklin Lakes, NJ, USA) using a FITC-labeled anti-human glycophorin monoclonal antibody (Dako, Denmark). Human erythrocytes were found to constitute 77.4% ± 19.9% (mean ± SD) of erythrocytes in mouse blood during periods of drug pressure. Mice were inoculated intravenously with 300 μL of 1% non-synchronized P. falciparum-infected huRBC. Follow-up of infection was performed by daily Giemsa-stained thin blood films drawn from the tail vein. In this paper, we report parasitemia as a percentage of all erythrocytes found in mouse peripheral blood; the true percentage of huRBC parasitized in the mice is higher, proportional to the level of chimerism, because murine erythrocytes cannot be infected but were included in counts.

Estimates of the total parasite biomass in each mouse were calculated based on the mean corpuscular volume of mouse erythrocytes (45 fL), the mean corpuscular volume of huRBC (86 fL), hematocrit in the mice of 0.7, weight of NSG mice (25 g), and a conservative estimate of 5.5 mL of blood per 100 g of mouse weight using the following equation:

Mice were initially infected with drug-naïve parasites from in vitro culture of cryopreserved stabilates and subsequently put under discontinuous sub-therapeutic AS drug pressure. Sodium AS (a gift from Sigma-Tau, Italy) was dissolved in 10% dimethyl sulfoxide (DMSO) in RPMI-1640 (stock solution 30 mg/mL) each day of injection, then diluted 10-fold in RPMI-1640, sterilized through a 0.22 μm Millex filter (Millipore, MA, USA), further diluted in sterile RPMI-1640 as appropriate, and delivered intravenously via the retro-orbital sinus.

For the single-dose protocol, one dose of AS (ranging from 2.4 mg/kg to 240 mg/kg) was given, then parasitemia was monitored every 24 h and allowed to recover back to pre-treatment levels (AS pressure cycle; APC) before a further dose of AS was administered. For the 2-day protocol, two doses of AS (starting at 2.4 mg/kg/injection up to 80 mg/kg/injection) were delivered 24 h apart, then parasitemia was monitored every 24 h and was allowed to recover back to pre-treatment levels (APC) before a further two doses of AS were given (i.e., for a 2-day dose of 2.4 mg/kg, the mouse was injected with a total of 4.8 mg/kg AS per APC). The length of APC varied from case to case. When parasitemia failed to drop significantly (see below) after exposure to a given dose, the concentration was increased. The parasite strain used for the 2-day protocol had already developed resistance to a single dose of 30 mg/kg AS, and was then subjected to the 2-day regimen starting at 2.4 mg/kg/injection. Parasite strains were named ART-R_x, where x is the dose of AS (in mg/kg) to which resistance was established in that strain.

To determine what should be considered a significant drop in parasitemia, the normal day-to-day fluctuation of parasitemia was calculated from 13 non-drug-exposed NSG-IV mice (geometric mean of variability ± 18.3%, 95% confidence interval (CI) 12.5–27%). Taken from this, the parasite was deemed to be resistant to a given dose when parasitemia failed to drop more than 27% by the next day (all reported measures of parasite reduction are from the day after drug administration). We analyzed the drop in parasitemia seen among five mice infected with the PAM-sensitive strain the day after a single administration of intravenous AS to define a ‘sensitive response’ to AS in this model. The mean reduction was 78.4% with a SD of 18.2%. We conservatively chose a drop in parasitemia greater than 60.2%, corresponding to the mean (1 SD) as the definition of a sensitive response to guide decisions about dosing. For definitive statistical comparisons of parasitemia responses, a paired t test was used. Stability of resistance was determined when required by re-challenging the parasite strain in its new host with the dose of drug to which it had last shown resistance. The ART-R P. falciparum strain was continuously perpetuated in vivo by sub-inoculation directly from one mouse to another by the intravenous route, except where otherwise indicated.

The primary technique used to determine IC₅₀ was the double-site enzyme-linked pLDH immunodetection assay, as previously described [39]. The ³H-hypoxanthine isotopic method [40] was used as a secondary confirmatory assay. All in vitro results shown below come from the double-site enzyme-linked pLDH immunodetection assay.

For both methods, P. falciparum parasites at 0.05% parasitemia, synchronized at ring stage, were incubated at 2% hematocrit in 96-well microtiter plates (Nunc, Sigma-Aldrich) with serial dilutions of various anti-malarial drugs in 200 μL of complete culture medium at 37 °C and 5% CO₂ for 72 h. Non-drug-exposed wells were used as positive controls, and wells containing non-infected huRBC served as negative controls.

Stock solutions of the drugs (5 mL,1.5 mg/mL) were prepared by dissolving sodium AS (gift from Sigma-Tau), chloroquine sulphate (Rhone-Poulenc-Rorer, Vitry, France), dihydroartemisinin (DHA; Sigma-Tau), pyrimethamine (ICN Biochemicals, Aurora, Ohio), quinine hydrochloride (Sanofi, Montpellier, France), lumefantrine (Sigma-Aldrich), and mefloquine hydrochloride (Hoffman-La Roche, Basel, Switzerland) in 10% DMSO in RPMI-1640, whereas amodiaquine dihydrochloride and halofantrine hydrochloride were dissolved in 30% DMSO in RPMI-1640. Drug solutions were diluted 10-fold in RPMI-1640, sterilized by filtration through a 0.22 μM filter, and serially diluted in a 96-well incubation plate.

IC₅₀ values were determined by performing a four-parameters, variable slope, non-linear regression analysis taking the least-squares fit without constraints, using Graph Pad Prism 6 software. Comparison of IC₅₀ values and hillslopes was performed using the extra sum-of-squares F test (GraphPad, Inc., CA, USA).

Mice infected with the ART-R₂₄₀ strain were given either single treatments or combinations of the following regimens: three doses of quinine hydrochloride 73 mg/kg every 8 h intravenously, four doses of halofantrine hydrochloride 1 mg/kg every 24 h intravenously, one dose of amodiaquine dihydrochloride 73 mg/kg orally (delivered by oro-gastric canula), one dose of chloroquine sulphate 73 mg/kg orally, or one dose of mefloquine hydrochloride 50 mg/kg intra-peritoneally, as previously described [41]. Stock solutions were made by dissolving 150 mg of quinine, chloroquine, and mefloquine in 5 mL of 10% DMSO, 150 mg of amodiaquine in 30% DMSO, and 60 mg of halofantrine in 30% DMSO, then dissolved 10-fold in RPMI-1640, and sterilized by filtration before being made up to the final concentration.

Plasma concentrations of AS and DHA in blood samples (40–60 μL) collected from the retro-orbital sinus in four mice at 1, 2, and 4 h post intravenous drug administration were determined by reversed phase liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) using an adaptation of the previously described method [42]. Murine plasma was purified by protein precipitation with acetonitrile, evaporation, and reconstitution in 10 mM ammonium formate/methanol (1:1) adjusted to pH 3.9 with formic acid. Separations were done on a 2.1 mm × 50 mm Atlantis dC18 3 μm analytical column (Waters, Milford, MA, USA). The chromatographic system (CTC Analytics AG, Zwingen, Switzerland) was coupled to a triple stage quadrupole Thermo Quantum Discovery Max mass spectrometer equipped with an electrospray ionization interface (Thermo Fischer Scientific Inc., Waltham, MA, USA). The selected mass transitions were m/z 221.1 → 163.1, with a collision energy of 14 eV for AS and DHA, and m/z 226.2 → 168.1, with a collision energy of 20 eV for the stable isotope-labeled internal standard DHA-¹³CD₄. Inter-assay precision obtained with plasma QC samples at 30, 300, and 3000 ng/mL of DHA and AS were 1.3, 2.1, 11.3%, and 7.3, 4.7, and 10.8%, respectively. Mean absolute deviation from nominal values of QC samples (30, 300, and 3000 ng/mL) during the analysis were 5.4, 5.9, and 1.3% and 3.8, 9.7, and 2.1%, for DHA and AS, respectively. The lower limit of quantification was 2 ng/mL. The laboratory participates in the External Quality Control program for anti-malarial drugs (http://​www.​wwarn.​org/​).

ART-R P. falciparum DNA was isolated from parasitized blood using QIAamp DNA mini kit (Qiagen, Limburg, Netherlands). A non-synonymous point mutation of ubp1 in P. chabaudi (PCHAS020720) was reported by others [43] as being a marker of ART resistance in a rodent model. The orthologous gene in P. falciparum (PF3D7_0104300) is conserved and was amplified using the primers (500 nM) forward: 5’-TACAGGCTTTATATAGTACAGTGTC-3′, reverse: 5’-TTTTCGTTCGTACTTATAGGCACAGG-3′, and AmpliTaq DNA Polymerase (1 U) (Hoffman-La Roche). The 451 bp PCR fragment was purified using the QIAquick PCR purification kit (Qiagen). Polymorphisms in PF3D7_0104300 were assessed by digesting the PCR fragment with the restriction enzymes Mae III for V3275F and Rsa I for V3306F, corresponding to V2697F and V2728F in PCHAS 020720, respectively.

Genes of interest in P. falciparum coding for the proteins RAD5, cNBP, RPB9, PK7, FP2A, Pfg27, Pfcrt, and Pfnhe, two fragments overlapping the kelch-13 propeller domain [44‐46], and Pfmdr1 gene were analyzed by PCR-sequencing. Primers used for Pfmdr1 PCR and sequencing were previously described by Basco and Ringwald [47], and Pfmdr1 gene copy analysis was performed as previously described [48]. For Pfnhe, two primer couples were designed for nested PCR on the basis of the 3D7 sequence. Control samples were taken from in vitro cultures of the P. falciparum 3D7 strain, and the sensitive progenitor PAM strain prior to any ART exposure (PAMwt); for the RAD5 experiment, additional control clinical isolates collected in the late 1990s were used from Brazil, Comoro Islands, Senegal, and Thailand. Experimental samples were recovered from P. falciparum-infected mice at various points during the ART resistance induction process (NSG415, 416, 424, 433, and 440). Genomic DNA was prepared using QIAamp DNA mini kit (Qiagen), according to the manufacturer’s instructions, in 50 μL of Milli-Q water; 1 μL of DNA was PCR-amplified with 500 nM of the corresponding forward and reverse primers (Additional file 1), 0.8 mM dNTPs, 1.5 mM MgCl₂, 2.5 U Taq DNA polymerase (Hoffman-La Roche) in a volume of 50 μL with the following cycling program: 2 min at 94 °C, 30 cycles of 15 s at 94 °C, 30 s at 57 °C, 45 s at 72 °C, and a final extension of 2 min at 72 °C. The total contents of the reaction were electrophoresed on a 1% agarose gel and stained with ethidium bromide. The amplicons were extracted from the gel using the QIAquick^® gel extraction kit (Qiagen). Concentration of the amplicons was measured by NanoDrop (Thermo Fischer Scientific Inc.) at 260 nm wavelength before sequencing of both strands was performed (Plateforme de séquençage, Institut Cochin, Paris/Eurofins MWG Operon). Sequences were analyzed with DNAstar software (DNAStar, Madison, WI, USA).

We infected seven mice with the PAM P. falciparum strain before any drug exposure to determine the LED. Single doses of 0.6, 0.3, and 0.15 mg/kg AS each caused a significant drop in parasitemia (> 27%, i.e., the upper 95% CI of normal fluctuation). Since 0.075 mg/kg AS failed to reduce parasitemia beyond normal day-to-day fluctuations, a single dose of 0.15 mg/kg AS (0.00375 mg AS/mouse) was established as the LED in this model (Fig. 1). Effective doses of AS produced pyknotic parasites as seen in humans (Additional file 2).

We applied intense, discontinuous, sub-curative AS drug pressure in vivo to high P. falciparum parasitemia in NSG mice using the intravenous route. After each drug exposure, parasitemia was allowed to recuperate back to pre-treatment levels (APC) and, once resistance was established, the AS dose was increased (Fig. 2 and Additional file 3). For the single-dose regimen, the median APC length was 4 days (range 2–14 days).

During the single-dose regimen, after pre-conditioning of the drug-naïve parasites with 3 single doses of AS in one mouse, we passed the parasite line through 7 generations of mice by sub-inoculation, using 5, 9, 6, 1, 6, 10, and 6 mice in each generation, respectively (total 43).

In the first generation, we let parasites multiply to high parasitemias (25–35%) creating a pool of ~ 1.3 × 10¹⁰P. falciparum-infected erythrocytes. We saw resistance to 2.4 mg/kg AS after 3 APC in 1 out of 4 mice exposed to that dose, then to 3.3 mg/kg AS after 2 APCs in 1 out of 3 mice, and to 4 mg/kg AS in 2 out of 2 mice exposed to a mean 1.5 APC.

In the second generation, resistance to 3.3 mg/kg was established in another mouse (1 APC), and to 4 mg/kg in 2 further mice (mean 1.5 APC, range 1–2). Later, we confirmed 4 mg/kg resistance in a new host. Seeing as resistance was so forthcoming, we increased drug pressure readily to 15 mg/kg AS, to which indeed 4 out of 5 mice exposed became resistant (mean 5 APC, range 2–9) (Additional file 4).

Resistance to 30 mg/kg AS then emerged in 2 out of 4 mice exposed to that dose (mean 1.5 APC, range 1–2). However, it was not stable and, in the third generation, an average of 3.6 APC (range 2–5) was required before it was re-established (ART-R₃₀). Subsequently, in 1 mouse, after applying variable-intensity drug pressure, resistance to 60 mg/kg AS was obtained (5 APC).

We confirmed the stability of resistance to 60 mg/kg AS (ART-R₆₀) immediately after sub-inoculation into the fourth generation, and after just three further exposures to 120 mg/kg AS, the strain showed the first signs of resistance to that dose.

In the fifth generation, we observed resistance to 120 mg/kg AS (ART-R₁₂₀) in all 4 mice exposed after an average of 3 APC (range 2–4). Then, in 1 mouse, the parasite went on to develop resistance against 240 mg/kg AS after 4 APC (78, 44, 60, and 13% reduction in parasitemia seen with each APC, respectively).

After sub-inoculation into the sixth generation, the parasite strain established resistance to 240 mg/kg AS in 4 out of 6 mice exposed to that dose (2.75 APC, 1–7).

In the seventh generation, resistance was immediately stable, after sub-inoculation, to 240 mg/kg AS in all 6 mice (ART-R₂₄₀) (mean ± SD percentage drop in parasitemia of sensitive control 78.4% ± 18.2% vs. ART-R₂₄₀ 9.1% ± 6.3%; p = 0.0002).

Since further dose doubling would exceed the lethal dose for 50% of mice [41, 49], 240 mg/kg was the highest dose administered. We used NSG mice infected with the sensitive progenitor PAM strain as controls, and all treatments using the above doses were found effective. This represents a 3200-fold decrease in in vivo AS sensitivity, occurring within 51 APC over a 45-week period (Table 1, Additional file 2, Additional file 3, and Additional file 5). Further, we observed gametocytes in thin blood smears from mice infected with parasites expressing the ART-R phenotype (Additional file 6).

Two doses of the same AS concentration administered 24 h apart – a double dose (DD) – caused a significant reduction in parasitemia in animals in which a single dose of the same concentration had failed.

We started with a concentration of 2.4 mg/kg/dose for the DD regimen using a parasite strain already resistant to a single dose of 30 mg/kg AS. The ART-R₃₀ strain became resistant to DD 2.4 mg/kg AS after just 1 APC. We passed the parasite line through four generations of mice with 6, 8, 3, and 4 mice in each generation, respectively. Once resistance was seen, we increased the dose concentration 2-fold, until reproducible resistance to DD 80 mg/kg AS (i.e., 160 mg/kg total) was achieved (ART-R_DD80) (Fig. 3, Additional file 7, and Additional file 8) (mean ± SD percentage drop parasitemia of sensitive control 95.9% ± 5.7% vs. ART-R_DD80 25.7% ± 0.6%; p = 0.03).

It was possible to select for resistance to the highest dose used in 41% of the mice that survived the 2-day protocol, in contrast with 80% of mice that underwent the single-dose protocol (Table 2).

We measured levels of AS and DHA at 1 and 2 h post injection of 120 mg/kg AS in four ART-R₁₂₀-infected mice (Additional file 9). Serum concentrations of DHA at 1 h were 3159, 3219, 1573, and 2423 ng/mL in each mouse, respectively, and we confirmed resistance to these levels on blood films drawn the following day. The mean t_1/2 of DHA in the infected NSG-IV model was 36 min (range 20.9–53.2 min).

Stability was assessed in three different manners:

We monitored 50% IC₅₀ values over the course of resistance development for both single-dose and 2-day regimens, and compared them to the sensitive progenitor.

The initial IC₅₀ (95% CI) values for the sensitive strain to AS and DHA were 10.7 nM (10.2–11.2) and 13.8 nM (12.9–14.6), respectively. The ART-R₃₀ strain did not show any increase in IC₅₀ for AS (11.5 nM (6.6–16.9); F = 0.525, p = 0.47); however, there was a significant change in the slope of the curve compared to the sensitive control (hillslope − 4.4 (–6 to –3.6) vs. − 1.9 (–6.4 to –0.8); F = 7.5, p = 0.008). It was not until the strain became resistant to 120 mg/kg AS in vivo that the IC₅₀ rose sharply for both AS (to 82.5 nM (69.5–95.8); F = 191.3, p < 0.0001) and DHA (to 54.6 nM (51.6–57.6); F = 300.3, p < 0.0001). The ART-R₂₄₀ strain reached an IC₅₀ of 100.3 nM (92.9–118.4) (F = 304.8, p < 0.0001) for AS.

In parasites submitted to a 2-day regimen, we saw the same pattern, with a delayed shift in IC₅₀ (Fig. 5 and Additional file 10).

Despite exclusive exposure to AS, the ART-R₂₄₀ parasite strain showed markedly decreased responses to quinine, amodiaquine, and halofantrine. Indeed, the IC₅₀ increased by 4.6-fold to quinine (49.7 nM (46.6–52.8) vs. 226.9 nM (145.8–392.1); F = 23.12, p < 0.0001), 3.8-fold to halofantrine (7.9 nM (7.3–8.6) vs. 30.4 nM (25.9–34.9); F = 159.3, p < 0.0001), and 11.7-fold to amodiaquine (11.3 nM (10.6–12.1) vs. 132.4 nM (5.5–149.3); F = 243.7, p < 0.0001); similarly, the DD ART-R_DD80 strain increased its IC₅₀ 2.1-fold to quinine (F = 98.9, p < 0.0001) and 4.5-fold to amodiaquine (F = 152.5, p < 0.0001). Sensitivities to chloroquine (50.1 nM (46.5–53.7) vs. 53 nM (42.7–68.3); F = 0.39, p = 0.54), mefloquine (41.7 nM (39.1–44.4) vs. 39.1 nM (34.1–44.5); F = 0.82, p = 0.37), lumefantrine (7.5 nM (6.3–8.7) vs. 7.8 nM (6.2–9.8); F = 0.13, p = 0.72), and pyrimethamine (16.2 nM (13.8–18.8) vs. 19.9 nM (16.5–24.6); F = 4.75, p = 0.05) remained unchanged (Fig. 6 and Additional file 10).

Since the model accommodates simultaneous in vitro and in vivo studies with P. falciparum, this pattern of in vitro co-resistance to main-stream anti-malarial drugs could also be analyzed in vivo (Fig. 7). Therapeutic doses of 219 mg/kg quinine did not induce any decrease in parasitemia in vivo using ART-R₂₄₀ strain (n = 4, mean ± SD percentage drop in parasitemia 4.8% ± 6.8%); the same dose was effective for the sensitive strain (n = 2, mean ± SD percentage drop in parasitemia 92.2% ± 0.01%; p = 0.03). In addition, we confirmed in vivo resistance to amodiaquine in 4 mice (mean ± SD percentage drop in parasitemia sensitive control 76.6% ± 5.2% vs. ART-R₂₄₀ 9.3% ± 0.14%; p = 0.03), and halofantrine in 3 mice (median, range percentage increase in parasitemia after 3 days of treatment 16.9%, 15.9–114.4%). Conversely, we observed in vivo susceptibility to treatment with mefloquine (2 mice, mean ± SD percentage drop in parasitemia 67.5% ± 7.8%; p = 0.005, compared to normal day-to-day fluctuation) and chloroquine (3 mice, mean ± SD percentage drop in parasitemia 73.3% ± 0.7%; p < 0.001, compared to normal day-to-day fluctuations).

We also addressed the in vivo response of ART-R₂₄₀ to two critical combinations in clinical use: AS plus amodiaquine was ineffective (mean ± SD percentage drop in parasitemia 13.1% ± 0.14% vs. 76.5% in the sensitive control), while AS plus mefloquine was effective (mean ± SD percentage drop in parasitemia 66.8% ± 33.6%; p = 0.004, compared to normal day-to-day fluctuations) (Fig. 7d, e).

Thus, in vivo findings mirrored the in vitro sensitivity profiles.

Restriction fragment length polymorphism assessment of two putative polymorphisms, V3275F and V3306F, in the P. falciparum orthologue of the ubp1 gene revealed no such mutation in the ART-R₂₄₀, ART-R_DD38.4, or parent PAM strain.

Genetic sequencing of PF3D7_1343400 (RAD5 homolog) encoding a putative DNA-repair protein identified the non-synonymous a3392t SNP (MAL13–1718319) in all of the ART-R P. falciparum samples recovered from experimental mice, wherein they had shown resistance to single doses of 38.4 mg/kg, 120 mg/kg, and 240 mg/kg AS, and to a 2-day regimen of 80 mg/kg/day AS. This RAD5 mutation was not identified in the wild type progenitor PAM strain prior to undergoing ART exposure (PAMwt), nor in any of four control clinical P. falciparum isolates collected from Brazil, Senegal, Comoro Islands, and Thailand in the late 1990s. We did not identify any mutation of cNBP in the PAMwt control or ART-R parasites (Additional file 11).

Sequencing of the putative Kelch-13 propeller domain in PF3D7_1343700 (kelch-13) showed no difference between control (3D7, PAMwt) and ART-R strains; it revealed none of the 20 non-synonymous SNPs that have been reported from clinical isolates, nor the SNP identified in P. falciparum that evolved in vitro ART tolerance (M476I) after being cultured for 5 years under artemisinin pressure [45] (Additional file 11). None of the other non-synonymous SNPs in RPB9, PK7, FP2a, or Pfg27 reported in association with the in vitro ART tolerance seen in that strain were found either [45]. Sequencing of exon two of PfCRT revealed the rare CVIKT haplotype [50] linked to moderate resistance to chloroquine in agreement with the in vitro response (chloroquine IC₅₀ = 53 nM).

Pfmdr1 analysis showed a duplication of gene copy number from 1 to 2 copies, and acquisition of the N86Y mutation after in vivo artemisinin drug pressure. No sequence changes were found in the 611 bp PfNHE fragment gene, flanking the DNNND repeat, which is related to quinine resistance [51].