Characterization of Plasmodium falciparum genes associated with drug resistance in Hodh Elgharbi, a malaria hotspot near Malian–Mauritanian border

A malaria hotspot in the southeastern region of Mauritania, near the Malian border, may hamper malaria control strategies. The objectives were to estimate the prevalence of genetic polymorphisms associated with drug resistance in Plasmodium falciparum isolates and establish baseline data. The study was conducted in two malaria-endemic areas in Hodh Elgharbi, situated in the Malian–Mauritanian border area. Blood samples were collected from symptomatic patients. Single nucleotide polymorphisms in Pfcrt, Pfmdr1, Pfdhfr, and Pfdhps were genotyped using PCR-restriction fragment length polymorphism, DNA sequencing and primer extension. The Pfmdr1 gene copy number was determined by real-time PCR. Of 280 P. falciparum-infected patients, 193 (68.9%) carried the Pfcrt 76T mutant allele. The Pfmdr1 86Y and 184F mutations were found in 61 (23.1%) of 264 isolates and 167 (67.6%) of 247 samples that were successfully genotyped, respectively. Pfmdr1 mutant alleles 1034C, 1042D and 1246Y were rarely observed. Of 102 P. falciparum isolates analysed, ten (9.8%) had more than one copy of Pfmdr1 gene. The prevalence of isolates harbouring at least triple mutant Pfdhfr 51I, 59R, 108 N/T was 42% (112/268), of which 42 (37.5%) had an additional Pfdhps 437G mutation. The Pfdhps 540E mutation was observed in four isolates (1.5%), including three associated with Pfdhfr triple mutant. Only two quintuple mutants (Pfdhfr-51I-59R-108N Pfdhps-437G-540E) were observed. The observed mutations in Pfdhfr, Pfdhps, Pfmdr1, and Pfcrt may jeopardize the future of seasonal malaria chemoprevention based on amodiaquine-sulfadoxine-pyrimethamine, intermittent preventive treatment for pregnant women using sulfadoxine-pyrimethamine, and treatment with artesunate-amodiaquine. Complementary studies should be carried out to document the distribution, origin and circulation of P. falciparum populations in this region and more widely in the country to assess the risk of the spread of resistance.

Ould Ahmedou Salem et al. Malar J (2017) 16:140 Background Before 2006, chloroquine and sulfadoxine-pyrimethamine were the first-and second-line drugs in Mauritania, respectively. The current anti-malarial treatment policy in Mauritania is based on artesunate-amodiaquine and artemether-lumefantrine as the first-and second-line treatment of uncomplicated malaria, respectively, regardless of Plasmodium spp. [1]. The efficacy and tolerance of artesunate-amodiaquine treatment have been confirmed in a recent clinical study conducted in southern Mauritania [2]. The Mauritanian Ministry of Health also recommends the use of sulfadoxine-pyrimethamine (SP) for intermittent preventive treatment to prevent malaria during pregnancy, particularly during the first and second pregnancies [3]. Due to the scarcity of clinical data on anti-malarial drug efficacy in Mauritania, molecular markers of drug resistance are convenient surrogate indicators to monitor and detect emergence of drug-resistant Plasmodium falciparum.
Dihydrofolate reductase (DHFR) and dihydropteroate synthase (DHPS) are known targets of pyrimethamine and sulfadoxine, respectively [4]. These drugs specifically inhibit the enzymes of the folate pathway to kill the parasites. P. falciparum chloroquine resistance transporter (Pfcrt) and P. falciparum multidrug resistance gene 1 (Pfmdr1) encode membrane transporters that have been associated with resistance to chloroquine and amodiaquine (Pfcrt) or chloroquine, amodiaquine, and amino-alcohols (mefloquine, lumefantrine; Pfmdr1). The mode of action of chloroquine and other quinoline-like drugs involves an interference with the plasmodial haem metabolism in the digestive vacuole. Chloroquine-resistant P. falciparum parasites either diminish the influx of chloroquine into the food vacuole or enhance efflux of chloroquine from the food vacuole, or both, resulting in a decreased drug accumulation within the food vacuole [5]. Distinct point mutations in Pfdhfr, Pfdhps and Pfcrt confer resistance to pyrimethamine, sulfadoxine and chloroquine/amodiaquine, respectively. Some studies have suggested that point mutations in Pfmdr1 are associated with resistance to quinoline-like drugs and artemisinin derivatives, but their role is not yet well established [6,7]. Moreover, it has been reported that treatment with artemether-lumefantrine selects for the pfmdr1 wild-type N86 allele [8][9][10]. Pfmdr1 amplification resulting in multiple gene copies is associated with resistance to amino alcohols (mefloquine and lumefantrine) [4,11]. Recent studies have demonstrated that mutations in Kelch propeller 13 is directly associated with artemisinin resistance [12].
Hodh Elgharbi region is one of the eight Mauritanian administrative regions where malaria transmission is seasonal, with the peak occurring in September and October.
Several studies have shown that P. falciparum is the predominant malaria species in this region and that 30-83% of febrile patients are infected with malaria parasites during the transmission season [2,[13][14][15][16]. In Hodh Elgharbi region, the northern Saharo-Sahelian area that borders the Saharan desert has been considered as malaria-free while the southern Sahelian part of the region is classified as holo-endemic [13]. The Malian-Mauritanian border is particularly vulnerable to the spread of malaria parasites with the influx of war refugees from the Malian side of the frontier. The problem of human population movement is further compounded by the presence of populations leading a nomadic lifestyle in this agropastoral area. The situation in Mali near the border area is quite similar in terms of malaria epidemiology and chloroquine resistance [17,18].
The objective of the present study was to compare the prevalence of Pfcrt, Pfmdr1, Pfdhfr, and Pfdhps mutations and the copy number of Pfmdr1 gene in P. falciparum isolates collected in two areas with different malaria endemicity in Hodh Elgharbi region after the introduction of artemisinin-based combination therapy (ACT) in 2006, in order to establish a baseline database for monitoring drug-resistant P. falciparum in the malaria hotspot in the Malian-Mauritanian border area.

Study area
The study was conducted in Hodh Elgharbi, which covers a surface area of 53,400 sq km and is composed of four districts (Aioun (the provincial capital), Kobeni, Tamchekett, Tintane). Kobeni city is located about 18 km from the Malian-Mauritanian border. Aioun and Tintane are located approximately 103 and 174 km to the north and northwest of Kobeni city, respectively. The population of Hodh Elgharbi is approximately 294,109 inhabitants [19]. The location and Saharo-Sahelian (Aioun and Tintane) and Sahelian (Kobeni) climates were described in an earlier study [16]. Human activities are dominated by livestock rearing (mainly cattle, sheep, and goats) and agriculture (mainly millet). There are five public health centres in Hodh Elgharbi, one health centre in each of its four districts and one hospital in Aioun.

Patients and blood sample collection
Between September and October 2010, corresponding to the peak season of malaria transmission, all febrile patients were enrolled after informed written consent. A rapid diagnostic test for malaria (SD Bioline P. falciparum histidine-rich protein II and Plasmodium vivax plasmodial lactate dehydrogenase antigen rapid diagnostic test; Standard Diagnostics, Inc, Yongin, Republic of Korea) was performed, and Giemsa-stained thin and thick blood films were prepared and examined under the microscope. Fifty µl of blood sample were spotted directly on Whatman 3MM filter paper, dried, and stored at room temperature for molecular analysis.

Genetic characterization of Plasmodium falciparum isolates
Total genomic DNA of each isolate, including human DNA, was extracted using the MaxMag ™ DNA Multi-Sample kit (Applied Biosystems, Warrington, UK) following the manufacturer's instructions. All molecular experiments were performed at the parasitology laboratory in Marseille, France in 2012-2013.

Pfmdr1 SNPs
Pfmdr1 (PF3D7_0523000) genotyping was performed by Sanger's method of DNA sequencing. Four pairs of primers were used to amplify two Pfmdr1 fragments carrying five polymorphisms associated with drug resistant phenotype (Table 1). N86Y and Y184F were on the first fragment MDR1-1 (590 base pairs, bp) and S1034C, N1042D and D1246Y on fragment MDR1-2 (968 bp). The reaction mixture, PCR conditions, amplicons purification and sequencing were described in previous studies [20,21].

Pfmdr1 gene copy number
Pfmdr1 copy number was determined using TaqMan real-time PCR (7900HT Fast Real-Time PCR system, Applied Biosystems, Courtaboeuf, France) using the single-copy gene β-tubulin (PF10_0084) as the reference. Each sample selected on the basis of a sufficient amount of DNA was genotyped in duplicate. The sequence of oligonucleotide primers and probes used, the preparation of reaction mixture, PCR conditions, and evaluation of PCR efficiency were described in detail in a previous study [21]. DNA extracted from the P. falciparum 3D7 reference clone, which has a single copy of each gene, was used as a calibrator, and β-tubulin housekeeping gene was used as a control in all experiments. The number of gene copy was determined using the 2 −ΔΔCt method [11].

Statistical analysis
Based on similar epidemiological and ecological characteristics in the study sites of Aioun and Tintane [16], data from these two districts were pooled and compared to those from Kobeni. The frequency of a particular mutant allele was calculated as the proportion of the specific mutant samples among the total number of samples successfully analysed for this mutation. Similarly, the frequencies of double, triple, quadruple, and quintuple mutants were determined as the proportion of subjects with two, three, four, or five mutations among the total numbers of samples tested. Pairwise comparison of mutation frequencies between the northern (Aioun and Tintane) and southern (Kobeni) sites was performed using the Chi square test. The distribution of Pfmdr1 and Pfcrt SNPs in relation to Pfmdr1 copy number was compared using Fisher's exact test. For statistical analyses, mixed alleles were considered to be mutant. The P value of ≤0.05 was considered as statistically significant.

Results
The clinical characteristics of the patients and parasitological features were published elsewhere [16]. In that study, laboratory diagnosis was based on microscopy and rapid diagnostic test. Plasmodium falciparum monoinfection was confirmed by PCR in 299 (146 in Kobeni, 92 in Tintane, and 61 in Aioun) patients. Plasmodium vivax was detected by PCR in 11 additional patients (nine in Kobeni, two in Tintane, none in Aioun).

Pfcrt polymorphisms
The key codon 76 was examined in the Pfcrt gene ( Table 2). The mutant allele 76T was identified in 193 of 280 (68.9%) samples that were successfully amplified, of which 160 (57.1%) were pure mutant allele and 33 (11.8%) had mixed alleles with both K76 and 76T. The difference in the proportions of mutant and mixed alleles in Kobeni and Tintane/Aioun was not statistically significant (P > 0.05).

Pfmdr1 copy number
Due to insufficient DNA left after characterizing four molecular markers, the copy number of Pfmdr1 was determined in 102 isolates (88 isolates from Kobeni and 14 from Tintane and Aioun) ( Table 6). All 102 isolates were successfully genotyped. Only one isolate (one of 102, 1.0%) was characterized with three copies of Pfmdr1.   Nine additional isolates (8.8%) had two copies of the gene. All isolates with multiple copies of Pfmdr1 were collected in Kobeni. All other isolates (n = 92) carried a single copy of Pfmdr1 gene. The copy number variations of Pfmdr1 were not associated with any mutation of this gene or with Pfcrt SNP (P > 0.05).

Discussion
The present study was conducted during the peak season of malaria transmission in three districts in the Hodh Elgharbi region. A high prevalence of mutant K76T Pfcrt allele was observed in the present study. This finding is in agreement with the previous clinical and molecular studies, confirming that Pfcrt is useful for the detection and surveillance of chloroquine-resistant P. falciparum in Hodh Elgharbi region [13]. The present study also revealed the high prevalence of Y184F and, to a lower extent, N86Y mutations in Pfmdr1 gene. It has been suggested that Pfmdr1 mutations may play a role in modulating the levels of resistance to several drugs [23]. Amodiaquine treatment failure has been associated with the selection of P. falciparum isolates carrying the mutant Pfcrt 76T allele and Pfmdr1 haplotype 86Y, Y184, and 1246Y [23][24][25][26]. In the present study, the majority of isolates (68.9% with either pure mutant or mixed alleles) had the mutant Pfcrt 76T allele, but the Pfmdr1 mutant haplotype associated with amodiaquine resistance was rare (for 1246Y mutant, two of 190 isolates, 1.0%). These molecular findings are in agreement with the high clinical efficacy of artesunate-amodiaquine in Kobeni (98.2% efficacy on day 28 after PCR correction) and the fact that amodiaquine monotherapy had not been used before adoption of ACT in the country [2]. An opposite trend for the selection of P. falciparum isolates with Pfmdr1 haplotype N86, 184F, and D1246 has also been reported after artemether-lumefantrine treatment [8][9][10]26]. Since artesunate-amodiaquine and artemether-lumefantrine are widely employed to treat P. falciparum in Mauritania, further surveillance of ACT resistance can be pursued using the combination of Pfcrt, Pfmdr1 and kelch 13 as surrogate markers for resistance to these drugs, in parallel with regular clinical evaluation of ACT efficacy. The limitations of the study on Pfmdr1 include the lack of complete molecular data for all isolates due to sub-optimal PCR conditions for the fragment spanning codons 1034, 1042, and 1246, the limited quantity of DNA available, and possible degradation of DNA due to poor storage conditions. The present study was conducted before the discovery of kelch 13, and insufficient blood sample did not allow further study on this novel marker.
Pfmdr1 gene copy number is pertinent for monitoring possible emergence of resistance to amino alcohols (lumefantrine and mefloquine), which is also influenced by Pfcrt and Pfmdr1 alleles [8][9][10]. At present, the proportion of isolates with increased gene copy number is limited in Mauritania, but further monitoring is required, in particular in Kobeni, where isolates with multiple copies of Pfmdr1 gene were found. In addition, an evaluation of the efficacy of artemether-lumefantrine in Kobeni is necessary to establish baseline clinical data.
Pfdhps A437G substitution is one of the components of the 'quintuple Pfdhps-Pfdhfr mutations' associated with SP resistance [27]. The analysis of molecular markers in the present study revealed the presence of 437G in 23% (35% including mixed alleles) of the isolates: 19% (38% including mixed alleles) in Aioun and 14% (24% including mixed alleles) in Kobeni. In an earlier study conducted in 1998, there were 22 and 16% of mutant 437G in Aioun and Kobeni, respectively [14]. There was only a slight increase in the proportion of mutant Pfdhps 437G allele in Kobeni after 12 years. By contrast, the proportion of mutant 437G allele almost doubled during the same period in Aioun. However, the other Pfdhps component of the 'quintuple mutant' , K540E, was observed in only four of 264 (1.5%) isolates in the present study and none in the 1998 study [14].
Triple Pfdhfr mutant (AIRNI) was present in a total of 106 of 269 (39%) isolates, either as pure alleles (n = 72, 27%) or mixed alleles (n = 34, 12.6%). In an earlier study, Eberl et al. reported 16.9% (ten of 59 isolates) and 12.6% (13 of 103 isolates) of triple Pfdhfr mutants in Aioun and Kobeni, respectively [14]. The proportions of triple Pfdhfr mutants increased two-to three-fold in Aioun and Kobeni between 1998 and 2010. This observation is consistent with the frequent use of SP for the treatment of uncomplicated malaria until 2006 and the continuous use of this drug for intermittent preventive treatment in pregnant women since 2006 [28]. The progression of Pfdhps mutants overtaking wild-type isolates between 1998 and 2010 seems to be slow in Kobeni compared to Tintane and Aioun where mutations in both Pfdhps and Pfdhfr occurred at a similar rate between 1998 and 2010. Quintuple mutant occurred rarely in the present study. Moreover, Pfdhfr I164L substitution, which confers a high level of pyrimethamine resistance, was found as mixed alleles in few isolates. The results of the molecular assays suggest that SP is still useful in targeted population (i.e., in pregnant women and possibly in infants) for intermittent preventive treatment. Other studies have suggested that intermittent preventive treatment with SP remains effective for fetal and maternal protection even in areas of high resistance to SP [29]. Current knowledge on malaria epidemiology is still inadequate to develop a concerted plan to control crossborder malaria in the study area. In addition to clinical and molecular studies on drug-resistant malaria and entomological surveys, sociological studies are needed to understand the patterns of human population movement of the nomads, refugees and local travellers along and across the Malian-Mauritanian border. These difficulties are compounded by the presence of P. vivax, which is notoriously difficult to eliminate due to hypnozoites, requiring primaquine for radical cure. One recent study failed to detect P. vivax along the Malian-Mauritanian border (Selibaby, Ould Yenge, Aioun, Kobeni, Timbedra, Nema) [30]. However, the results of other studies, including the present study, indicate the presence of P. vivax in this hotspot [15,16,31]. A coordinated regional malaria control programme will be required in the effort to control cross-border malaria that involves both P. falciparum and P. vivax [32].