Molecular studies have demonstrated that mutations in the Plasmodium falciparum chloroquine resistance transporter gene (Pfcrt) play a major role in chloroquine resistance, while mutations in P. falciparum multidrug resistance gene (Pfmdr-1) act as modulator. In Madagascar, the high rate of chloroquine treatment failure (44%) appears disconnected from the overall level of in vitro CQ susceptibility (prevalence of CQ-resistant parasites <5%) or Pfcrt mutant isolates (<1%), strongly contrasting with sub-Saharan African countries. Previous studies showed a high frequency of Pfmdr-1 mutant parasites (>60% of isolates), but did not explore their association with P. falciparum chloroquine resistance. To document the association of Pfmdr-1 alleles with chloroquine resistance in Madagascar, 249 P. falciparum samples collected from patients enrolled in a chloroquine in vivo efficacy study were genotyped in Pfcrt/Pfmdr-1 genes as well as the estimation of the Pfmdr-1 copy number. Except 2 isolates, all samples displayed a wild-type Pfcrt allele without Pfmdr-1 amplification. Chloroquine treatment failures were significantly associated with Pfmdr-1 86Y mutant codon (OR = 4.6). The cumulative incidence of recurrence of patients carrying the Pfmdr-1 86Y mutation at day 0 (21 days) was shorter than patients carrying Pfmdr-1 86N wild type codon (28 days). In an independent set of 90 selected isolates, in vitro susceptibility to chloroquine was not associated with Pfmdr-1 polymorphisms. Analysis of two microsatellites flanking Pfmdr-1 allele showed that mutations occurred on multiple genetic backgrounds. In Madagascar, Pfmdr-1 polymorphism is associated with late chloroquine clinical failures and unrelated with in vitro susceptibility or Pfcrt genotype. These results highlight the limits of the current in vitro tests routinely used to monitor CQ drug resistance in this unique context. Gaining insight about the mechanisms that regulate polymorphism in Pfmdr1 remains important, particularly regarding the evolution and spread of Pfmdr-1 alleles in P. falciparum populations under changing drug pressure which may have important consequences in terms of antimalarial use management.
Citation: Andriantsoanirina V, Ratsimbasoa A, Bouchier C, Tichit M, Jahevitra M, Rabearimanana S, et al. (2010) Chloroquine Clinical Failures in P. falciparum Malaria Are Associated with Mutant Pfmdr-1, Not Pfcrt in Madagascar. PLoS ONE 5(10): e13281. doi:10.1371/journal.pone.0013281
Editor: Georges Snounou, Université Pierre et Marie Curie, France
Received: June 22, 2010; Accepted: September 12, 2010; Published: October 13, 2010
Copyright: © 2010 Andriantsoanirina et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was supported by grants from the Institut de Médecine et d'Emiologie Appliquée (IMEA), Fondation Léon MBA, Paris, France and the Genomics Platform, Pasteur Génopole, Pasteur Institute, France. Samples collection in Madagascar was funded by the Global Fund to Fight AIDS, Tuberculosis and Malaria, round 3 (Community Action to Roll Back Malaria, grant no. MDG-304-G05-M). Valerie Andriantsoanirina is a graduate PhD student funded by the Institut Pasteur de Madagascar (Bourse “Girard”) and the Direction des Affaires Internationales (Institut Pasteur). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Plasmodium falciparum resistance to chloroquine (CQ) has emerged at least from six independent foci (South East Asia, Venezuela, Colombia, Papua New Guinea, India and Philippines) in the late 1950s and in the 1960s , , . Molecular evolutionary studies have demonstrated that P. falciparum CQ-resistant parasites from South-East Asia have entered in East Africa (Kenya and Tanzania) in the late 1970s and spread across the African continent within two decades . P. falciparum-related deaths rose sharply after the spread of CQ-resistant parasites in sub-Saharan Africa, affecting mostly children under 5 years of age , , .
CQ-resistance in P. falciparum shows some biological similarities with the multiple drug resistance phenotype of mammalian tumour cells, as both involve expulsion of drug out of the cytosol of the cell and can be reversed by calcium channel antagonists such as verapamil . Based on molecular allele exchange studies and analysis of genetic crosses, it is today generally accepted that the major role in CQ resistance is determined by polymorphisms in Pfcrt, a gene encoding a transporter which promotes, in its mutated forms, drug efflux from the parasite digestive vacuole, while Pfmdr-1 modulates the level of in vitro CQ-resistance , , . In addition, mutations or amplifications of Pfmdr-1 gene can play a significant role in P. falciparum resistance to diverse antimalarials such as mefloquine, quinine or artemisinin derivatives . Specific combinations of Pfcrt and Pfmdr-1 alleles, resulting in varying responses to CQ (and amodiaquine), appeared geographically restricted, which may explain why some field studies reported an association between Pfmdr-1 polymorphisms and CQ resistance and other studies did not .
In Madagascar, CQ was first introduced in 1945 and was largely used during more than 50 years, in particular in a campaign for malaria prevention treatment in children . A shortage in CQ supply in the 1970s was followed by a large-scale epidemic in the late 1980s. Wide-scale access to CQ treatment resumed afterwards, and CQ is still used today though its replacement by the artesunate plus amodiaquine combination therapy is recommended as first-line treatment for uncomplicated cases since 2005. The first clinical cases of CQ resistance were suspected in 1975. In 1981, other cases of treatment failures were reported and in vitro tests showed isolates with high 50% inhibitory concentration for CQ (IC50) indicative of in vitro CQ-resistance. However, during the 1981-2008 period, several studies outlined lower rates of in vitro CQ-resistance and therapeutic failures in Madagascar than in other African countries . Currently, the prevalence of in vitro CQ resistance does not exceed 5% of isolates and Pfcrt mutant parasites are found in less than 1% of collected isolates . Surprisingly, these figures appear disconnected from the high level of clinical treatment failures (44%) , most of which (∼90%) are late clinical/parasitological treatment failures. Interestingly, previous surveys outlined the presence of Pfmdr-1 mutations in more than 60% of isolates, suggesting that Pfmdr-1 mutants could be responsible for CQ clinical failure in Madagascar.
The aim of the present work was to document the association of Pfmdr-1 with in vivo CQ clinical failure and in vitro CQ resistance. To this end, two independent sets of P. falciparum samples collected in 2006–2007 from in vivo and in vitro susceptibility studies were genotyped for Pfcrt and Pfmdr1 genes as well as the estimation of the Pfmdr-1 copy number. Isolates that meet the inclusion criteria were included in the association analysis (240 from in vivo study and 90 from in vitro testing). In addition, the evolutionary dynamics of the Pfmdr-1 locus in the Malagasy parasite populations were also assessed. Our data show that in this particular setting, Pfmdr-1 polymorphism on multiple genetic backgrounds, in absence of mutations in Pfcrt gene or Pfmdr-1 gene amplification was significantly associated with late chloroquine clinical failures and unrelated with the overall level of in vitro CQ susceptibility, raising the limits of the current in vitro tests routinely used to monitor CQ drug resistance.
Materials and Methods
P. falciparum samples from eight sentinel sites involved in the monitoring of the antimalarial drug resistance in Madagascar  were studied. The collection sites were located in the four main malarious epidemiological strata: Ejeda and Ihosy in the South (sub-desert stratum, epidemic prone), Maevatanana and Miandrivazo in the West (tropical stratum, seasonal and endemic area), Tsiroanomandidy and Moramanga in the foothills of the Central Highlands (highlands stratum, low-endemic area) and Farafangana and Andapa in the East (equatorial stratum, perennial endemic area). Malaria transmission, estimated by annual entomological inoculation rate (EIR, number of bites of infected anophelines per person sleeping indoors), varied according to the sites from 0.2 (in the South) to 240 (in the East) . The prevalence of anti-MSP1 antibodies used as proxy of the burden of malaria in each site was previously determined at the same collection period  and were included in our analysis.
Patients and in vivo study
In 2006–2007, multi-site prospective chemotherapeutic studies were carried out for assessing the therapeutic efficacy of antimalarial therapies recommended by the National Malaria Control Programme (NMCP) (registration number ISRCTN36517335) . Patients with microscopically confirmed uncomplicated P. falciparum malaria were randomized into one of the four treatment groups: Chloroquine (CQ), Amodiaquine (AQ), Sulfadoxine-Pyrimethamine association (SP) and Artesunate-Amodiaquine combination (ASAQ). According to the 2003 WHO protocol, patients aged from six months to 15 years, presenting with P. falciparum monoinfection (parasitaemia from 1,000 to 200,000/µL), axillary temperature ≥37.5°C, body weight >5 kg, haemoglobin (Hb) ≥5 g/dL, without severe malnutrition, signs of severity, or a concomitant disease were included in the clinical trial. Once written informed consent was given by the patient or the guardian and after randomization, patients were administered either CQ (10 mg/kg on days 0 and 1, and 5 mg/kg on day 2), AQ (10 mg/kg on days 0, 1, and 2), SP (25 mg/kg sulfadoxine and 1.25 mg/kg pyrimethamine as a single dose on day 0) or ASAQ (AS: 4 mg/kg on days 0, 1, and 2 and AQ:10 mg/kg on days 0, 1, and 2) and seen on days 1, 2, 3, 7, 14, 21 and 28, or any intervening day if they were unwell for malaria infection to assess the clinical and parasitological efficacy of the drug regimens. Blood samples were obtained by finger prick on the enrolment and all follow-up days including any unscheduled day to use for analysis of thick and thin blood smears and for storage on filter paper for further molecular analysis. Haemoglobin concentration was measured on day 0 using a HemoCue haemoglobinometre (HemoCue AB, Ängelholm, Sweden).
Collection of P. falciparum isolates for in vitro drug sensitivity testing
As part of the surveillance of antimalarial drug resistance in Madagascar, fresh P. falciparum clinical isolates were routinely collected in 2006–2007 from symptomatic patients prior to treatment in six of the eight sites (Ihosy, Maevatanana, Miandrivazo, Tsiroanomandidy, Moramanga and Farafangana). Venous blood samples were collected in tubes coated with EDTA (Vacutainer tubes, Becton Dickinson, Rutherford, NJ, USA), from malaria-positive patients (>2 years old) after they gave their consent for participation in the study (approval number 007/SANPF/2007). Presence of malaria parasites was evaluated by using a rapid diagnostic test based on the detection of Plasmodium-specific lactate dehydrogenase (pLDH) (OptiMAL-IT, DiaMed AG, Cressier sur Morat, Switzerland). Positive patients were treated with the artesunate-amodiaquine combination, according to the National Malaria Control Programme (NMCP), and samples were sent to the Malaria Research Unit of the Institut Pasteur of Madagascar in a controlled cool box at 4°C. Drug sensitivity assays were done using the classical isotopic 48-h test, as described , on blood samples with parasitaemia ≥0.1%, available within 48 h after blood collection and from patients declaring no antimalarial drug intake during the previous 7 days. For CQ sensitivity assay, three control wells without drug were used as control and each concentration ranging from 12.5 to 1,600 nM was studied in duplicate or triplicate. The IC50, i.e. the drug concentration corresponding to 50% of 3H-hypoxanthine uptake by the parasites in drug-free control wells, was determined by probit/logit regression analysis. The quality of the assays was controlled by using P. falciparum reference lines (3D7 Africa, CQ-sensitive clone and FCM29 Cameroon, CQ-resistant clone).
Parasite DNA was available from whole blood or capillary blood transferred to filter paper (Whatman, Maidstone, UK). DNA was extracted from blood spots with Instagene® Matrix resin (BioRad©, Marnes la Coquette, France), according to the manufacturer's instructions or directly from 100 µl of whole blood by using the phenol-chloroform method . Parasite species was confirmed by using real-time polymerase chain reaction as described by de Monbrison et al. .
Pfmdr-1 and Pfcrt genotyping in isolates from day 0 and day of recurrence
Pfmdr-1 and Pfcrt genes were amplified in samples from enrolment (and day of recurrence in cases of CQ treatment failure) by nested PCR approach, as described . For Pfmdr-1 gene, two fragments (549 bp, from codons 23 to 206 and 938 bp, from codons 966 to 1278) and for Pfcrt gene, three fragments (450 bp, from codon 37 in exon 2 to codon 125 in exon 3, 447 bp, from codon 121 in exon 3 to codon 211 in exon 4 and 850 bp, from codon 199 in exon 3 to codon 316 in exon 7) were sequenced. Sequencing reactions were carried out with ABI Prism BigDye Terminator cycle sequencing ready reaction kit and were run on a model 3730 xl genetic analyzer (Applied Biosystems, Courtaboeuf, France). Electrophoregrams were visualized and analyzed with CEQ2000 genetic analysis system software (Beckman Coulter, Villepinte, France) and amino acid sequences were compared with the wild-type amino acid sequences (GenBank accession numbers, AF030694 for Pfcrt, and XM_001351751 for Pfmdr-1). The presence of single nucleotide polymorphisms (SNPs) was confirmed by reading both strands. Sequences of insufficient quality were either resequenced or rejected. Pfmdr-1 and Pfcrt alleles were reconstructed from full sequences presenting an unambiguous single allele signal at each nucleotide position.
Pfmdr-1 copy number determination
The Pfmdr-1 copy number was measured by TaqMan real-time PCR (Rotor-Gene 6000, Corbett Research, Sydney, Australia) relative to the single copy of the ß-tubulin (used as a housekeeping gene), as described . All samples were run in duplicate in 25 µL reaction mixtures. For each run, the Pfmdr-1 copy number was measured relative to the numbers in two standard calibrator parasite clonal lines (3D7 Africa and Dd2), by the ΔΔCT method . Reference DNA clone W2 (which has three copies of Pfmdr-1) was used as the quality control in each run.
Analysis microsatellite flanking Pfmdr-1
Two microsatellite markers (MS loci 956456 and 957861), located on chromosome 5 and extending ∼10 kb downstream the Pfmdr-1 gene were used to determine the evolutionary history of Pfmdr-1 alleles. Microsatellite polymorphism was analyzed using a nested PCR strategy, as previously described by Mehlotra et al. . Microsatellite PCR products were genotyped on the basis of size, using a GeneScan 500 LIZ size standard on an ABI Prism 3730 XL DNA analyzer.
Assessment of isolate clonality
The number of genotypes present in isolates collected from in vivo and in vitro assays, was estimated by using an allelic family-specific nested PCR (MAD20, K1, and RO33 for Pfmsp-1 and 3D7 Africa and FC27 for Pfmsp-2) . Clonality was defined as the highest number of alleles detected at either of the two loci and used to classify isolates as monoclonal or polyclonal and to distinguish recrudescence from new infection for all patients failing therapy after day 7 (isolates from day 0 and day of recurrence). All PCR amplifications contained a positive control (genomic DNA from strains W2, HB3, and 3D7 Africa) and a negative control (no target DNA).
Data were entered and verified using Microsoft Excel© software, and analyzed using EpiInfo 6.04© software (Centers for Disease Control and Prevention, Atlanta, GA, United States) and XLSTAT© for Windows XP (Addinsoft, Paris, France). For in vivo clinical trial samples, criteria of inclusion for the analysis were (i) patients with complete 28-days follow-up, (ii) absence of reinfection, (iii) wild-type Pfcrt alleles in isolates collected at day 0 and day of recurrence and (iv) successful determination of Pfmdr-1 alleles in isolates from day 0 and day of recurrence. For in vitro testing samples, criteria of inclusion for the analysis were (i) monoclonal isolates by using msp1/msp2 genotyping, (ii) wild-type Pfcrt alleles and (iii) successful determination of Pfmdr-1 alleles.
The Mann-Whitney U test or Kruskal-Wallis method were used for non-parametric comparisons, and Student's t test or one-way analysis of variance for parametric comparisons. For categorical variables, Chi-squared or Fisher's exact tests were used to assess significant differences in proportions. Odds ratios (OR) and their 95% confidence intervals (95%CI), describing the association between CQ clinical outcomes and exposure variables were determined by conditional logistic regression.
For the multivariate analysis, variables with P-values <0.25 were initially introduced into the model and removed following a backwards-stepwise selection procedure to leave only those with a P-value <0.05 in the final model. The relation between Pfmdr-1 86Y mutation in isolates from day of recurrence and clinical response to CQ treatment was assessed by survival analysis using the Kaplan-Meier method and the log-rank test. The median asexual parasite densities clearance time (in percent of value on day 0) was also compared between patients according to the Pfmdr-1 alleles (86N and 86Y) found at day 0 (and day of recurrence). All reported P-values are two-sided and were considered statistically significant if less than 0.05.
Genetic diversity was assessed by Nei's unbiased expected heterozygosity (He) from haploid data and calculated as He = [n/(n−1)][1−pi] (n = the number of isolates sampled; pi = the frequency of the ith allele . Population genetic differentiation was measured using Wright's F statistics ; population genetic parameters were computed with FSTAT software, v2.9.4 .
The study protocol was reviewed and approved by the Ethics Committee of the Ministry of Health of Madagascar (approval number 007/SANPF/2007; registration number ISRCTN36517335). Informed written consent was provided by all patients or their parents/guardians before inclusion in the study.
In vivo and in vitro sample collection
In the in vivo clinical trial, a total of 8363 febrile children attending in health centres from the sentinel sites were screened for falciparum malaria and 1873 were microscopically positive for P. falciparum. They were assigned in different treatment groups. Among the 320 patients treated with CQ, 240 of them (75%) met the study analysis criteria. According to the WHO 2003 protocol, 119 patients were classified as cured and 121 as failing treatment (13 patients failed before day 7, 20 patients between days 8–14, 46 patients between days 15–21 and 42 patients between days 22–28). Late treatment failures accounted for 89.3% of the overall treatment failures. Details are given in figure 1.
Flowchart of patients: Enrollment, Randomization, Follow-up, Molecular genotyping.
Among the 420 P. falciparum isolates tested for in vitro susceptibility to CQ, 372 (88.6%) were successfully assayed. All isolates presented a wild type Pfcrt allele; ninety of them, meeting the study analysis criteria, were included in the analysis. Among the 90 isolates in vitro tested, the geometric mean and the median of CQ IC50 was 18.7 nM (95%CI 14.7–23.7 nM) and 22 nM (95%CI 16.2–25.0 nM), respectively.
Pfmdr-1 alleles, copy number of Pfmdr-1 gene
Pfmdr-1 alleles were determined for the 240 isolates from the in vivo efficacy study (day 0 and day of recurrence, when available) and for 90 isolates for which the IC50 for CQ was determined. Amongst in vivo isolates, three of five major SNPs previously related to CQ-resistance were observed at day 0, N86Y (147/240, 61.2%), Y184F (172/240, 71.7%), and D1246Y (80/240, 33.3%), present in eight different alleles: the wild-type allele NYD (11.2%), three single-mutant alleles (NFD, 22.9%; YYD, 14.6% and NYY, 1.2%), three double-mutant alleles (YFD, 17.9%; NFY, 3.3% and YYY, 1.2%) and one triple-mutant allele (YFY, 27.5%). Discordant alleles were found in 24 isolates between isolate from day 0 (wild-type NYD) and isolate from day of recurrence (13 single-mutant alleles YYD, 8 single-mutant alleles NFD and 3 double-mutant alleles YFD). There was a non-random association between the N86Y and D1246Y loci (R2 = 0.23, P = 0.004) and F184Y and D1246Y loci (R2 = 0.28, P = 0.0003).
Amongst in vitro isolates, only three isolates had an IC50>100 nM: one triple-mutant allele (YFY, IC50 = 142 nM), one double-mutant allele (YFD, IC50 = 126 nM) and one single-mutant allele (NFD, IC50 = 140 nM). Comparison of geometric mean IC50 between isolates harbouring different Pfmdr-1 alleles showed no significant difference: 18.0 nM (n = 37, 95%CI 12.1–26.6 nM) in Pfmdr-1 86N isolates compared to 19.2 nM (n = 53, 95%CI 14.1–26.0 nM) in Pfmdr-1 86Y isolates, 14.3 nM (n = 72, 95%CI 8.2–25.0 nM) in Pfmdr-1 184Y isolates compared to 20.0 nM (n = 18, 95%CI 15.3–26.1 nM) in Pfmdr-1 184F isolates and 21.4 nM (n = 61, 95%CI 13.5–36.7 nM) in Pfmdr-1 1246D isolates compared to 18.6 nM (n = 21, 95%CI 12.3–28.3 nM) in Pfmdr-1 1246Y isolates.
The Pfmdr-1 copy number was determined for 290 isolates (219 from in vivo studies and 71 from in vitro assays). The mean copy number was 0.93 (ranging from 0.74 to 1.33). When the value was rounded to the nearest integer, no Pfmdr-1 amplification was observed.
Pfmdr-1 microsatellite polymorphism
The polymorphisms in two microsatellite loci flanking the wild-type coding sequence and mutant-type alleles (single-mutants to triple-mutant) are shown in Figure 2. The MS locus 956456 had 6 alleles (ranging from 210 to 228 bp) and the MS locus 957861 had 4 alleles (ranging from 179 to 185 bp), displaying 7 different haplotypes. Microsatellite markers were moderately polymorphic for wild- and mutant-type alleles with a mean Nei's unbiased expected heterozygosity (He) ranging from 0.26 (NFY) to 0.66 (NYD). The Wright's fixation index analysis showed a significant absence of genetic differentiation between allele populations.
Distribution and prevalence of the two microsatellite loci flanking Pfmdr-1 gene (MS 956456, panel A and MS 957861, panel B) in 53 Plasmodium falciparum isolates collected from Madagascar in 2006–2007.
In vivo outcomes and risk factors associated to CQ treatment failure
To test whether any collected variables (ecological environment, patients and isolates characteristics) were associated with clinical response to CQ treatment, we compared them between both cured- and failing treatment patient groups (Table 1). Among these, four variables were significantly associated with the risk of CQ treatment failure: (i) prevalence of anti-PfMSP-1 antibodies previously estimated in the site (P = 0.04, 1.8-fold increased risk of CQ treatment failures in area of unstable malaria), (ii) age group (P<0.0001, increased risk of CQ clinical failures inversely correlated to age: 1.6-fold in the 6–10 years age group and 4.5-fold in the 0.5–5 years age group), (iii) haemoglobin concentration at day 0 (P<0.0002, 2.4-fold increased risk of CQ treatment failures for people with haemoglobin concentration <10 g/dL at day 0) and (iv) presence of the Pfmdr-1 86Y mutation in isolates collected on day 0 (P = 0.009, 2-fold increased risk of CQ treatment failures) or on day of recurrence (P<0.0001, 4.2-fold increased risk of CQ treatment failures).
In the multivariate analysis, variables with P-values <0.25 were initially introduced into the model (prevalence of anti-PfMSP-1 antibodies, gender, age, haemoglobin concentration at day 0, multiplicity of infection, presence of Pfmdr-1 codon mutation at position 86 in isolates from day 0 and day of recurrence and at position 184 in isolates from day of recurrence). Following a backwards-stepwise selection procedure, CQ treatment failures were significantly associated with presence of Pfmdr-1 codon mutation at position 86 in the parasites collected on day of recrudescence (OR = 4.6, 95%CI 2.3 to 8.9, P<0.0001), age (OR = 1.2, 95%CI 1.1 to 1.3, P = 0.0002) and prevalence of antiPfMSP-1 antibodies (OR = 1.02, 95%CI 1.0 to 1.05, P = 0.02).
Asexual parasite clearance following CQ-treatment displayed significant differences according to Pfmdr-1 allelic form at position 86 at day 1 (Pfmdr-1 86N: 77.1%, n = 89 and Pfmdr-1 86Y: 95.3%, n = 151, P = 0.0001), day 7 (Pfmdr-1 86N: 3.7%, n = 85 and Pfmdr-1 86Y: 34.3%, n = 142, P<0.0001) and day 14 (Pfmdr-1 86N: 2.0%, n = 78 and Pfmdr-1 86Y: 54.9%, n = 129, P<0.0001) (Figure 3).
Asexual parasites curves following CQ-treatment according Pfmdr-1 mutation at position 86 in isolates from patients included in the analysis, Madagascar in 2006–2007.
Cumulative incidence of recurrence of patients and Pfmdr-1 86 mutation
The relation between Pfmdr-1 86Y mutation in isolates from day 0 (or day of recrudescence, if available) and clinical response to CQ treatment by using the time of recurrence, showed a markedly difference between the two curves (Figure 4). The cumulative incidence of recurrence of patients carrying the Pfmdr-1 86Y allele (n = 163, median time of recurrence 21 days) was significantly shorter than patients carrying Pfmdr-1 86N (n = 77, median time of recurrence 28 days). The log-rank test shows that the two curves differ significantly (P<0.0001) and the Pfmdr-1 N86Y allele had a significant influence on the time of recurrence. The hazard ratio comparing the hazards in the two groups was estimated at 0.36 (95%IC: 0.23–055).
Kaplan-Meier curves of cumulative incidence of recurrence of patients over the 28-day follow-up period according Pfmdr-1 mutation at position 86 in isolates of day of recurrence from patients included in the analysis, Madagascar in 2006–2007.
In the isolates from Madagascar studied here which harboured a wild type Pfcrt, Pfmdr-1 polymorphism, in the absence of any noticeable gene amplification, seems to play a major role in late chloroquine clinical failures without affecting the overall level of in vitro CQ susceptibility. This situation contrasts with findings from multiple settings across the African continent, where CQ resistance depends primarily on mutations in Pfcrt gene and on additional mutations in Pfmdr-1 gene, which may increase the level of resistance afforded by Pfcrt. Pfcrt mutant alleles seem universally found, albeit with different haplotypes in different geographical regions . The association of Pfmdr-1 with CQ resistance was not found in some areas , possibly reflecting the fact that the impact of the Pfcrt/Pfmdr-1 combination mutation depends on the genetic background of the strain ,  as demonstrated by studies with genetically manipulated lines or recombinant progeny of experimental crosses ,  and the history of use of the antimalarial drugs (chloroquine and quinine being the two mains drugs used in Madagascar).
The observation that parasites with a wild Pfcrt allele were associated with a high rate of therapeutic failures of chloroquine is totally new. Several hypothesis may account for these findings. First, a particular Pfcrt haplotype, restricted to the Malagasy area and different from previously described resistant haplotypes, was present but not detected during this study. As we sequenced for all isolates a large part of Pfcrt gene including five codons consistently associated with CQ resistance in other settings, this hypothesis appears unlikely. It also is worth noting that so far the rare Pfcrt mutant alleles found in two sites from Madagascar (Andapa and Tsiroanomandidy) had a CVIET or CVIDT haplotype  and most likely were imported from the neighbouring Comoros Islands. Why did the CVIET haplotype not spread largely across Madagascar in spite of a significant drug pressure, as it did in Asia and Africa and even in the neighbouring Comoros Islands, remains to explain. A second hypothesis is that another gene, distinct from Pfcrt and not identified at present, is the main responsible for CQ resistance in Madagascar. Though this hypothesis cannot be formally ruled out, it appears also unlikely in view of the observed, strong association with Pfmdr1. A third possibility, is that the Pfcrt allele is present at a low fraction in the time 0 sample, too low to be detected (masked by the wild type allele and not detected by the PCR followed by DNA sequencing methodology) and too low to translate into a shifted IC50 in vitro (as these parasites represent a low fraction, they incorporate a low amount of 3H hypoxanthine and remain unnoticed). We also think this possibility unlikely, as no mutant Pfcrt could be detected in the samples collected from patients with late therapeutic failures. Indeed, the seminal studies by Djimde et al. in Mali showed that the minority CQ resistant clone selected during a treatment and causing a therapeutic failure accounted for a substantial proportion of the parasites on the day of recurrence . Juliano et al. , using a nonradioactive heteroduplex tracking assay, reported recently that in a unique series of 17 patients in Madagascar, two (11.7%) harboured Pfcrt CVIET-resistant haplotype. The proportion of mutant parasites was 1.7 and 2.9% of the total parasite recurrent population obtained 14 days after the onset of CQ treatment, and neither of the two patients harboured any detectable CVIET parasites prior to treatment. In both cases the “recurrent” isolate 14 days after treatment by CQ contained a vast majority of CQ susceptible parasites (98.3% and 97.1% respectively), indicating that parasite genetic loci other than Pfcrt came into play in these recrudescences. We did not detect any Pfcrt resistant haplotype on the recurrence day of the 121 therapeutic failures of our series. Whether this reflects a somewhat lower sensitivity of our methodology compared to Juliano et al remains to be determined, but it is certainly consistent with the conclusion that these recrudescences are not associated with selection of a minority, mutant Pfcrt clone. The fourth hypothesis is that mutant alleles of Pfmdr-1 play a major role in CQ clinical outcome. Ten years ago, transfection methods were used to explore the role of Pfmdr-1 mutations in CQ resistance . Introduction of wild-type polymorphisms into the resistant 7G8 line resulted in reduction of CQ resistance but introduction of mutations in the susceptible D10 line did not confer CQ resistance, which suggested that Pfmdr-1 was not sufficient by itself to confer resistance. Lessons drawn from a recent genetic cross however indicate that observations with one or the other P. falciparum line are not directly applicable to all lines . Different phenotypes of CQ resistance (and amodiaquine resistance) exist in the world, differing in different geographical areas. The contribution of Pfmdr-1 mutations in CQ resistance varies depending on the particular Pfcrt haplotype with which they are associated .
The situation with regard to CQ resistance in Madagascar is unique. There were high rates of therapeutic failures, but those failures occurred in close to 90% of cases more than 7 days after treatment (Late Treatment Failures, LTFs) whereas in sub-Saharan Africa the proportion of LTFs is usually near 20% , , , , . The proportion of isolates with detectable Pfcrt mutant parasites was totally disconnected from the rate of clinical failures. Moreover, no selection of Pfcrt mutant haplotypes between the first day of treatment and the day of recurrence was observed, which contrasted strongly with studies conducted in sub-Saharan countries . We found also a selection of the N86Y or Y184F mutant alleles between the first day of treatment and the day of recurrence (10%), suggesting that Pfmdr-1 has a major role in CQ clinical failures in Madagascar. However, this particular type of resistance was not reliably detected by the classical isotopic 48-h test used in our study, as only 3.3% of isolates showed phenotypic resistance (IC50>100 nM). The follow-up of circulating parasite densities over time in recurrent malaria episodes, showing rapid (but not complete) asexual parasites clearance within 72-hours following first day of treatment (Figure 3), was consistent with the apparent in vitro susceptibility of the isolates obtained in therapeutic failure cases. These results highlight the limits of the in vitro tests routinely used to monitor the antimalarial drug resistance. Based on the detection of the “trophozoites to schizonts” growth during a single parasite erythrocytic cycle, it appears that such in vitro assays can accurately detect high levels of resistance (corresponding usually to early treatment failures) but are unable to evaluate phenotypic resistance yielded by slow acting molecular mechanisms (or unusual mechanism), likely related to late failure treatments , .
The analysis of microsatellites flanking Pfmdr-1 showed that mutations occurred on multiple genetic backgrounds (Figure 2). As previously described by Mehlotra et al. , we did not observe a strong reduction of heterozygosity in the CQ resistant parasite population, underlining that drug pressure from CQ treatment exerts a weaker selection on Pfmdr-1 than on Pfcrt. This “soft” selection of Pfmdr-1 mutant alleles may also be due to quinine pressure, as quinine was for a long time and until recently the antimalarial most commonly prescribed by practitioners in Madagascar . The spread of Pfmdr-1 mutant alleles may reflect a better fitness of parasites harbouring them or a better efficacy of the transmission .
Apart from the Pfmdr-1N86Y allele, the main risk factors associated with CQ therapeutic failures, lower age and lower prevalence of anti-PfMSP-1 antibodies, were inversely correlated with the immune status of patients. These results are similar with what has been reported for Pfcrt mutants in Mali and other countries where increasing age, reflecting the progressive acquisition of partial immunity, helped in the clearance of resistant parasites . Hence, the immune status of patients probably explains why a significant proportion of patients having isolates harbouring a mutant Pfmdr-1 allele were successfully cured by CQ in our series.
At present, CQ is being replaced by newer artemisinin-based combination drugs, such as artesunate-amodiaquine as first-line treatment in uncomplicated falciparum malaria in Madagascar. As Pfmdr-1 is a major modulator of resistance to these drugs , monitoring in the future in which measure the high proportion of Pfmdr-1 mutants observed in Madagascar can have an impact on the resistance of these antimalarials, even if at the moment this impact seems modest . More generally, gaining insight about the mechanisms that regulate the genetic variation at Pfmdr1 is important, particularly regarding the evolution and spread of Pfmdr-1 alleles in P. falciparum populations under changing drug pressure which may have important consequences in terms of antimalarial use management.
We thank the patients and healthcare workers involved in the national network for the surveillance of malaria resistance in Madagascar (Réseau d'Etude de la Résistance, RER) from which these samples were obtained, and the staff of the Ministry of Health of Madagascar for their collaboration.
Conceived and designed the experiments: AR OMP RD DM. Performed the experiments: VA MT MJ SR RR. Analyzed the data: VA CB OMP RD DM. Contributed reagents/materials/analysis tools: AR CB OMP DM. Wrote the paper: VA OMP RD DM.
- 1. Talisuna AO, Bloland P, D'Alessandro U (2004) History, dynamics, and public health importance of malaria parasite resistance. Clin Microbiol Rev 17: 235–254.
- 2. Wellems TE, Hayton K, Fairhurst RM (2009) The impact of malaria parasitism: from corpuscles to communities. J Clin Invest 119: 2496–2505.
- 3. Wootton JC, Feng X, Ferdig MT, Cooper RA, Mu J, et al. (2002) Genetic diversity and chloroquine selective sweeps in Plasmodium falciparum. Nature 418: 320–323.
- 4. Rowe AK, Rowe SY, Snow RW, Korenromp EL, Schellenberg JR, et al. (2006) The burden of malaria mortality among African children in the year 2000. Int J Epidemiol 35: 691–704.
- 5. Snow RW, Trape JF, Marsh K (2001) The past, present and future of childhood malaria mortality in Africa. Trends Parasitol 17: 593–597.
- 6. Trape JF, Pison G, Preziosi MP, Enel C, Desgrees du Lou A, et al. (1998) Impact of chloroquine resistance on malaria mortality. C R Acad Sci III 321: 689–697.
- 7. Krogstad DJ, Gluzman IY, Kyle DE, Oduola AM, Martin SK, et al. (1987) Efflux of chloroquine from Plasmodium falciparum: mechanism of chloroquine resistance. Science 238: 1283–1285.
- 8. Fidock DA, Nomura T, Talley AK, Cooper RA, Dzekunov SM, et al. (2000) Mutations in the P. falciparum digestive vacuole transmembrane protein PfCRT and evidence for their role in chloroquine resistance. Mol Cell 6: 861–871.
- 9. Mehlotra RK, Mattera G, Bockarie MJ, Maguire JD, Baird JK, et al. (2008) Discordant patterns of genetic variation at two chloroquine resistance loci in worldwide populations of the malaria parasite Plasmodium falciparum. Antimicrob Agents Chemother 52: 2212–2222.
- 10. Duraisingh MT, Cowman AF (2005) Contribution of the pfmdr1 gene to antimalarial drug-resistance. Acta Trop 94: 181–190.
- 11. Andriantsoanirina V, Menard D, Tuseo L, Durand R (2010) History and current status of Plasmodium falciparum antimalarial drug resistance in Madagascar. Scand J Infect Dis 42: 22–32.
- 12. Andriantsoanirina V, Ratsimbasoa A, Bouchier C, Jahevitra M, Rabearimanana S, et al. (2009) Plasmodium falciparum drug resistance in Madagascar: facing the spread of unusual pfdhfr and pfmdr-1 haplotypes and the decrease of dihydroartemisinin susceptibility. Antimicrob Agents Chemother 53: 4588–4597.
- 13. Menard D, Ratsimbasoa A, Randrianarivelojosia M, Rabarijaona LP, Raharimalala L, et al. (2008) Assessment of the efficacy of antimalarial drugs recommended by the National Malaria Control Programme in Madagascar: up-dated baseline data from randomized and multi-site clinical trials. Malar J 7: 55.
- 14. Mouchet J, Carnevale P, Coosemans M, Julvez J, Manguin S, et al. (2004) Biodiversité du paludisme dans le monde. Paris: John Libbey Eurotext. 428 p.
- 15. Razakandrainibe R, Thonier V, Ratsimbasoa A, Rakotomalala E, Ravaoarisoa E, et al. (2009) Epidemiological situation of malaria in Madagascar: baseline data for monitoring the impact of malaria control programmes using serological markers. Acta Trop 111: 160–167.
- 16. Rakotonirina H, Barnadas C, Raherijafy R, Andrianantenaina H, Ratsimbasoa A, et al. (2008) Accuracy and reliability of malaria diagnostic techniques for guiding febrile outpatient treatment in malaria-endemic countries. Am J Trop Med Hyg 78: 217–221.
- 17. de Monbrison F, Angei C, Staal A, Kaiser K, Picot S (2003) Simultaneous identification of the four human Plasmodium species and quantification of Plasmodium DNA load in human blood by real-time polymerase chain reaction. Trans R Soc Trop Med Hyg 97: 387–390.
- 18. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25: 402–408.
- 19. Nei M (1978) Estimation of Average Heterozygosity and Genetic Distance from a Small Number of Individuals. Genetics 89: 583–590.
- 20. Wright S (1965) The interpretation of population structure by F-statistics with special regard to systems of mating. Evolution 19: 395–420.
- 21. Goudet J (1995) FSTAT (Version 1.2): A Computer Program to Calculate F-Statistics. J Hered 485–486.
- 22. Mita T, Tanabe K, Kita K (2009) Spread and evolution of Plasmodium falciparum drug resistance. Parasitol Int 58: 201–209.
- 23. Valderramos SG, Valderramos JC, Musset L, Purcell LA, Mercereau-Puijalon O, et al. (2010) Identification of a mutant PfCRT-mediated chloroquine tolerance phenotype in Plasmodium falciparum. PLoS Pathog 6: e1000887.
- 24. Sa JM, Twu O, Hayton K, Reyes S, Fay MP, et al. (2009) Geographic patterns of Plasmodium falciparum drug resistance distinguished by differential responses to amodiaquine and chloroquine. Proc Natl Acad Sci U S A 106: 18883–18889.
- 25. Randrianarivelojosia M, Fidock DA, Belmonte O, Valderramos SG, Mercereau-Puijalon O, et al. (2006) First evidence of pfcrt mutant Plasmodium falciparum in Madagascar. Trans R Soc Trop Med Hyg 100: 826–830.
- 26. Djimde A, Doumbo OK, Cortese JF, Kayentao K, Doumbo S, et al. (2001) A molecular marker for chloroquine-resistant falciparum malaria. N Engl J Med 344: 257–263.
- 27. Juliano JJ, Randrianarivelojosia M, Ramarosandratana B, Ariey F, Mwapasa V, et al. (2009) Nonradioactive heteroduplex tracking assay for the detection of minority-variant chloroquine-resistant Plasmodium falciparum in Madagascar. Malar J 8: 47.
- 28. Reed MB, Saliba KJ, Caruana SR, Kirk K, Cowman AF (2000) Pgh1 modulates sensitivity and resistance to multiple antimalarials in Plasmodium falciparum. Nature 403: 906–909.
- 29. Kazadi WM, Vong S, Makina BN, Mantshumba JC, Kabuya W, et al. (2003) Assessing the efficacy of chloroquine and sulfadoxine-pyrimethamine for treatment of uncomplicated Plasmodium falciparum malaria in the Democratic Republic of Congo. Trop Med Int Health 8: 868–875.
- 30. Menard D, Madji N, Manirakiza A, Djalle D, Koula MR, et al. (2005) Efficacy of chloroquine, amodiaquine, sulfadoxine-pyrimethamine, chloroquine-sulfadoxine-pyrimethamine combination, and amodiaquine-sulfadoxine-pyrimethamine combination in Central African children with noncomplicated malaria. Am J Trop Med Hyg 72: 581–585.
- 31. Ogungbamigbe TO, Ojurongbe O, Ogunro PS, Okanlawon BM, Kolawole SO (2008) Chloroquine resistant Plasmodium falciparum malaria in Osogbo Nigeria: efficacy of amodiaquine+sulfadoxine-pyrimethamine and chloroquine+chlorpheniramine for treatment. Mem Inst Oswaldo Cruz 103: 79–84.
- 32. Ogwang S, Engl M, Vigl M, Kollaritsch H, Wiedermann G, et al. (2003) Clinical and parasitological response of Plasmodium falciparum to chloroquine and sulfadoxine/pyrimethamine in rural Uganda. Wien Klin Wochenschr 115: Suppl 345–49.
- 33. Sirima SB, Tiono AB, Konate A, Diarra A, Castelli F, et al. (2003) Efficacy of artesunate plus chloroquine for the treatment of uncomplicated malaria in children in Burkina Faso: a double-blind, randomized, controlled trial. Trans R Soc Trop Med Hyg 97: 345–349.
- 34. Wein S, Maynadier M, Tran Van Ba C, Cerdan R, Peyrottes S, et al. (2010) Reliability of antimalarial sensitivity tests depends on drug mechanisms of action. J Clin Microbiol 48: 1651–1660.
- 35. Sutherland CJ, Alloueche A, Curtis J, Drakeley CJ, Ord R, et al. (2002) Gambian children successfully treated with chloroquine can harbor and transmit Plasmodium falciparum gametocytes carrying resistance genes. Am J Trop Med Hyg 67: 578–585.