Prevalence of molecular markers of anti-malarial drug resistance in Plasmodium vivax and Plasmodium falciparum in two districts of Nepal

Background Sulphadoxine-pyrimethamine (SP) and chloroquine (CQ) have been used in treatment of falciparum and vivax malaria in Nepal. Recently, resistance to both drugs have necessitated a change towards artemisinin combination therapy (ACT) against Plasmodium falciparum in highly endemic areas. However, SP is still used against P. falciparum infections in low endemic areas while CQ is used in suspected cases in areas with lack of diagnostic facilities. This study examines the prevalence of molecular markers of CQ and SP resistance in P. falciparum and Plasmodium vivax to determine if high levels of in vivo resistance are reflected at molecular level as well. Methods Finger prick blood samples (n = 189) were collected from malaria positive patients from two high endemic districts and analysed for single nucleotide polymorphisms (SNPs) in the resistance related genes of P. falciparum and P. vivax for CQ (Pfcrt, Pfmdr1, Pvmdr1) and SP (Pfdhfr, Pfdhps, Pvdhfr), using various PCR-based methods. Results and discussion Positive P. vivax and P. falciparum infections were identified by PCR in 92 and 41 samples respectively. However, some of these were negative in subsequent PCRs. Based on a few P. falciparum samples, the molecular level of CQ resistance in P. falciparum was high since nearly all parasites had the Pfcrt mutant haplotypes CVIET (55%) or SVMNT (42%), though frequency of the Pfmdr1 wild type haplotype was relatively low (35%). Molecular level of SP resistance in P. falciparum was found to be high. The most prevalent Pfdhfr haplotype was double mutant CNRNI (91%), while frequency of Pfdhps double mutant SGEAA and AGEAA were 38% and 33% respectively. Combined, the frequency of quadruple mutations (CNRNI-SGEAA/AGEAA) was 63%. Based on P. vivax samples, low CQ and SP resistance were most likely due to low prevalence of Pvmdr1 Y976F mutation (5%) and absence of triple/quadruple mutations in Pvdhfr. Conclusions Based on the limited number of samples, prevalence of CQ and SP resistance at molecular levels in the population in the study area were determined as high in P. falciparum and low in P. vivax. Therefore, CQ could still be used in the treatment of P. vivax infections, but this remains to be tested in vivo while the change to ACT for P. falciparum seems justified.


Background
Malaria is one of the major health problems in South-East Asia where 1.3 billion people (76% of total population) are at risk, causing around 120,000 deaths yearly [1]. Plasmodium vivax malaria is the most widespread and prevalent in this region but large knowledge gaps still exits that preventing the assessment of the technical feasibility of its elimination still exist [2]. In Nepal, roughly 80% (22.5 million) of the population lives in malaria endemic areas of which seven million reside in highly endemic areas [3]. P. vivax is the predominant species that causes around 80-90% of the total malaria cases while Plasmodium falciparum is the main cause of malaria epidemics in Nepal [4]. Malaria transmission may occur throughout the year but mostly from March to November with peaks in June, July and August [5].
Chloroquine (CQ) is the first-line drug for the treatment of P. vivax infections as well as for the treatment of suspected P. falciparum infections in situations where diagnosis is unavailable; sulphadoxine-pyrimethamine (SP) has been the drug of choice for laboratory confirmed uncomplicated P. falciparum infections. Recently, artemisinin combination therapy (ACT) was introduced in 13 high endemic areas, replacing SP, as treatment for laboratory confirmed uncomplicated P. falciparum due to reports of high SP resistance [4].
Anti-malarial drugs have been one of the most important tools in the control of malaria over the last 50 years [6,7]. However, a major impediment to this strategy is the capacity of the parasites to develop drug resistance. This resistance has spread almost globally with regard to CQ and SP [7,8]. In Nepal, CQ resistance to P. falciparum was first observed in 1984, then in 1986, 1987 and 1988 [9]. CQ resistance in P. vivax has not been recorded. SP resistance in P. falciparum was first recorded in 1996 and then again in 2003 [9,10]. These studies were based on reports of treatment failures. Molecular markers of drug resistance, which are now commonly applied as an adjunct in the surveillance of anti-malarial drug resistance, have had only been used limited in Nepal. The advantage of using molecular markers of drug resistance is that the level of drug resistance can be studied retrospectively, such as after a drug has been abandoned due to reports of high levels of drug resistance in vivo.
In Nepal, a limited study of molecular markers of antimalarial drug resistance only in P. falciparum has been carried out [26]. Therefore, in light of the changing drug policy in Nepal towards using ACT; artemetherlumefantrine (AL) in some areas for P. falciparum and to fill some of the gaps in the knowledge of resistance to commonly used anti-malarials, the present study investigated the prevalence of molecular markers of anti-malarial drug resistance in both species by analysing the SNPs of Pfcrt, Pfmdr1, Pfdhfr and Pfdhps in P. falciparum and Pvmdr1 and Pvdhfr in P. vivax using samples from patients living in two highly endemic districts of Nepal.

Sample collection
Ethical clearance was obtained from the Nepal Health Research Council, Kathmandu, for the study (Ref. No. 266). After obtaining informed consent from either the patient or legal guardian, blood samples were collected from patients attending private or government hospitals or health posts in Jhapa (n = 146) and Banke (n = 43) districts in Eastern and Western Nepal, respectively ( Figure 1). Malaria is highly endemic in both districts and are two of the 13 districts where ACT (AL) was implemented in 2004 [4]. These districts share borders with the Indian states of Uttar Pradesh, Bihar and Assam where malaria is endemic [5]. Finger prick blood samples were collected either on Whatman no. 3 filter paper or as blood smear slides from malaria positive patients confirmed by microscopy or rapid diagnostic test. where annual malaria transmission rate are ≥ 1/1,000 and 0-1/1,000 population respectively [5,53]. The figure has been modified from [54] to provide more detail.

Genetic characterization of the parasites
Extraction of DNA from filter paper and blood smear slides were done by chelex method as described in [27] and [28] respectively. Plasmodium species (either vivax or falciparum) were differentiated by nested polymerase chain reaction (PCR) as in [29] and identification of Plasmodium malariae and Plasmodium ovale were done as previously [30]. Analysis of SNPs at Pvmdr1 was done by sequencing the nested PCR products as in [31]. Analysis of SNPs at Pvdhfr was performed by sequence specific oligonucleotide probes-enzyme-linked immunosorbent assay (SSOP-ELISA) as in [32] and for Pfcrt, Pfdhfr and Pfdhps, analysis were done by SSOP-ELISA as previously described in [33]. Finally, Pfmdr1 was analysed by PCR-restriction fragment length polymorphism (RFLP) as described elsewhere [34,35].

Results
Of the 189 samples, PCR identified 113 samples (60%) positive for Plasmodium, of which 72 (64%) and 21 (19%) samples were P. vivax and P. falciparum respectively while 20 (17%) of the samples carried mixed P. falciparum/P. vivax infections (Table 1). No P. malariae or P. ovale infections were detected. Positive samples from Banke were few (11%, n = 12) when compared to Jhapa (89%, n = 101) and was pooled since no apparent differences were observed between the sampling sites. The species PCR positive samples were further categorized according to type of infection, gender and age group. The highest prevalence of malaria was found in males (85%, n = 96) and the prevalence of infections were highest in the age group of 21-40 years (66%, n = 75) followed by 0-20 years (18%, n = 20) ( Table 1).

Prevalence of SNPs in Pfcrt
The prevalence of haplotypes at c72-76 could be determined for 31 out of 41 P. falciparum positive samples including mixed haplotypes (n = 2). The mutant haplotypes CVIET and SVMNT were identified in 55% (n = 17) and 42% (n = 13) of samples respectively while 10% samples (n = 3) contained the wild type CVMNK ( Figure 2).

Prevalence of SNPs in Pvmdr1
Only a fraction of samples (n = 47) were sequenced by alignment of DNA sequence with the wild type Salvador 1 as reference strain and out of these, 39 samples had satisfactory results. Only a fragment of 784 bp covering the codons 958, 976 and 1076 of the Pvmdr1 gene had been sequenced. A novel mutation was identified at T958M in all samples while F1076L was observed in 37 samples (95%) and Y976F was only identified in two samples (5%). Including the T958M mutation in constructing haplotypes for c958, c976 and c1076, the most frequent haplotype was the double mutant haplotype MYL (82%, n = 32), whereas the single mutant MYF and the triple mutant MFL haplotypes were determined in 13% (n = 5) and 5% (n = 2) of the cases, respectively ( Figure 6). Finally, novel SNPs were observed in two samples; one sample was found to contain a TTT to TCT mutation at c979 (F979S), and an ATG to GTG mutation at c980 (M980V). Further, one sample contained an AGT to AAT mutation at c1080 (S1080N).

Discussion
Several therapeutic efficacy studies were carried out in Nepal to determine the efficacy of currently used   anti-malarial drugs. For P. falciparum, studies done from 1984 to 1990 had shown 38% in vivo resistance to CQ. Similarly, SP efficacy studies conducted between 1996 to 2000 showed 57-73% treatment failures in P. falciparum infections [36]. Contrarily, a trial performed in 2005 in Jhapa showed only 12% SP treatment failure [26]. In Nepal, changes in anti-malarial drug policies have been based on declining efficacy of first line drugs when the therapeutic efficacy fell below 90% [4]. In line with this, AL was introduced against P. falciparum in 2004 in 13 highly endemic districts. SP is still the first line of treatment in the other 52 endemic districts, however, ACT will gradually replace SP in the near future [4].
This study determined CQ and SP resistance at molecular level to explore whether the high level of in vivo CQ and SP resistance previously found in Nepal resulting in the recent policy change, is reflected at molecular level as well. Polymorphisms related to CQ resistance in P. falciparum were determined in both Pfcrt and Pfmdr1. Combined, the Pfcrt mutant haplotypes CVIET (55%) and SVMNT (42%) were determined in a total of 97% of samples indicating high levels of CQ resistance. These haplotypes were also the most common haplotype determined from the Indian state of Assam (33% CVIET and 62% SVMNT) in 2006 [37].
Studies done in Malawi [38] and Tanzania [39] had shown the re-emergence of CQ sensitive P. falciparum populations after the withdrawal of CQ as a first line drug. However, this was not observed here, most probably due to the current anti-malarial drug policy of Nepal which recommends CQ for laboratory unconfirmed P. falciparum infection [4]. Regarding Pfmdr-1, this study found a high prevalence of wild types N86 (67%), Y184 (56%) and D1246 (100%). Further, 35% of the constructed haplotypes were the haplotype NFD. This haplotype has previously been implied as a marker of emerging AL tolerance [35,40]. Therefore, it is important to continue the monitoring of this particular marker in Nepal now with AL as the first-line treatment in highly endemic areas.
Studies have indicated that Y976F is the most important marker of CQ resistance and F1076L of lesser importance [11,31,41]. The analysis of Pvmdr1 identified a novel mutation at T958M, which has only been reported in samples from Madagascar [42] and its role in CQ resistance is unknown. Likewise, mutation at Y976F was determined only in two samples (5%). This suggests the presence of limited prevalence of CQ resistant haplotypes in P. vivax population. Finally, SNPs at c979, c980 and c1080 were observed, which have not been described elsewhere and the implications of these are unknown.
Regarding SP resistance in P. falciparum, SNPs at c51, c59 and c108 of Pfdhfr were identified. The most frequent mutant haplotype was the double mutated CNRNI (91%), which was also the most frequent haplotype determined at 64% from Jhapa in 2005 [26]. Thus, it seems that the prevalence of double mutant haplotype has increased. Furthermore, in 2003/2004, the CNRNI haplotype was the most frequent haplotype determined at 50-91% and 74-80% from Utter Pradesh and Assam in India, respectively [43,44]. Likewise, the triple mutated CIRNI was detected at low prevalence (6%) in the current study and in concordance with the previous study done in Jhapa (prevalence of 10% and 3% in 2002 and 2005 respectively) [26]. Further, CIRNI was also determined low in samples from Utter Pradesh (5%) and  Assam (10%) in 2003 [43,44], which indicates a low but consistent prevalence of the triple mutant haplotypes in the region. For Pfdhps, SGEAA (38%) and AGEAA (33%), the haplotypes that have been associated with treatment failure of SP [45] were the most prevalent in this study. In the study done in Assam in 2003/2004, SGEAA and AGEAA were determined in 2-6% and 11-17% of the samples, respectively [43,44]. This may suggest an increase of sulphadoxine resistant haplotypes in the region.
For Pfdhfr and Pfdhps combined, both quadruple (double mutations in Pfdhps and Pfdhps) and quintuple mutations (triple mutations in Pfdhfr and double mutations in Pfdhps) have been associated with treatment failures [43,46,47]. The present study determined quadruple mutations in 63% of samples. Further, the low prevalence (17%) of CNRNI-SGEAA/AGEAA in Assam in 2003 [43,44] may point towards the increase of prevalence of quadruple mutations in the region. Thus, this high prevalence may imply a high risk of SP treatment failure, which is in concordance with two previous studies done in Jhapa that recorded 12% and 21% treatment failures of SP [26,48].
Although SP is not used deliberately against P. vivax infections in Nepal, P. vivax is exposed to SP through mixed infections and/or incorrect diagnosis [20][21][22]. Regarding SNPs in the Pvdhfr, the double mutant FR2TN which is common in Uttar Pradesh (34%) and Assam (50%) [49], was the major mutant haplotype in the current study (47%) as well. The triple and quadruple Pvdhfr mutant haplotypes associated with high levels of in vivo SP resistance [50] were not found, indicating limited levels of P. vivax SP resistance in the study area.
The SNP patterns observed in Pfcrt, Pfdhfr and Pvdhfr genes in Nepal are similar to the studies performed in India, especially in the states that share borders with Nepal [37,43,49]. This indicates that either the parasites are under similar drug pressure in both countries or the free movement of people for work and other purposes between these countries are responsible for carrying parasites with similar drug resistance profiles. In this study, high percentages (85%) of the malaria positive were identified in males. Similar results were determined from Jhapa in 2005 (73%) and 2007 (75%) [26,51]. A study done in 2008 showed that around 37% of patients, particularly the males, had a history of a recent visit to India [52]. Although this study did not evaluate this aspect, it may be the possible cause for the higher percentage of infection in males and thus the similar patterns of SNPs in the resistance genes of P. falciparum and P. vivax.
A limitation to the present study is the small sample size. The majority of the samples collected on glass slides were found negative by PCR, which may be due to low parasitemia or misdiagnosis by microscopy or simply poor success in DNA extraction. Thus, the results should be interpreted with caution.

Conclusions
This study has determined low molecular prevalence of resistance to CQ and SP in P. vivax and a high level of CQ and SP resistance in P. falciparum in a limited number of samples from two districts in Nepal. Therefore, CQ is most likely still effective in the treatment of P. vivax infections, but this remains to be tested in vivo while the change to ACT against P. falciparum infection seems justified at molecular level.