Falciparum malaria in the north of Laos: the occurrence and implications of the Plasmodium falciparum chloroquine resistance transporter (pfcrt) gene haplotype SVMNT
Tomas Jelinek (corresponding author), Sabine Dittrich and Jörg M Stohrer, Institute of Tropical Medicine, Spandauer Damm 130, 14050 Berlin, Germany. Tel.: +49 30 30116 701; Fax: +49 30 30116 888; E-mail: email@example.com
Michael Alifrangis and Insaf Khalil, Centre for Medical Parasitology at Institute of Medical Microbiology and Immunology and Institute of Public Health, University of Copenhagen, Denmark.
Vonthalom Thongpaseuth, Viengxay Vanisaveth, Rattanaxay Phetsouvanh and Samlane Phompida, Centre of Malariology, Parasitology and Entomology, Vientiane, Laos.
Objective The Pfcrt-gene encodes a transmembrane protein located in the Plasmodium falciparum digestive vacuole. Chloroquine resistant (CQR) strains of African and Southeast Asian origin carry the Pfcrt-haplotype (c72–76) CVIET, whereas most South American and Papua New Guinean CQR stains carry the SVMNT haplotype.
Method Eighty-eight samples from an area with reported in vivo Chloroquine and in vitro Amodiaquine-resistance were screened for the K76T mutation and their Pfcrt-haplotype (c72–76) using a new SSOP-ELISA.
Results Hundred percent of the analysed samples showed the K76T mutation which is highly associated with in vivo drug failure. This very high rate of a CQR-marker is alarming in an area were CQ is still used as first line drug. The distribution of the three main Pfcrt-haplotypes was as follows: 68% CVIET, 31% SVMNT, 0% CVMNT.
Conclusions These data show, for the first time, the South American/PNG -haplotype (SVMNT) on mainland Southeast Asia. SVMNT-haplotype and others might be associated with a decreased efficacy of Amodiaquine and could therefore be potential markers for of amodiaquine resistance (AQR). If there is a correlation between AQR and the SVMNT-haplotype as suggested, 31% prevalence of a potential resistance marker is cause for concern.
Since the first report of chloroquine (CQ) resistant falciparum malaria in Southeast Asia and South America almost half a century ago, drug resistance has posed a major problem in malaria control. Today, CQ resistance (CQR) occurs almost everywhere where Plasmodium falciparum does.
CQ is still the first line-treatment recommended by the Lao authorities, although different studies show a resistance level as high as 46% in some areas of Laos (Mayxay et al. 2003). CQ and other blood schizontocides target haemoglobin digestion within the digestive vacuole or lysosome in the growing erythrocytic stages of the parasites. The resistance process is not yet fully understood but it is believed to follow changes in lysosomal integral membrane proteins (Fidock et al. 2000). It is widely believed that the most crucial mutations for CQR are non-silent mutations in the Pfcrt gene (Fidock et al. 2000). The most important mutation is the amino acid change at position K76T (Fidock et al. 2000). Recent studies investigated not only the single mutation at codon 76 but codons from 72 to 76, and found three main haplotypes: CVMNK, CVIET and SVMNK (Wootton et al. 2002). It was proposed that the different haplotypes belong to different geographic regions. The CVMNK haplotype as the worldwide identical wild type, CVIET belonging to most of Africa/Asia and SVMNT as the South American/Papua New Guinea (PNG) allele (Wootton et al. 2002). Additional to these three main haplotypes various other amino patterns were reported in individual samples [SVMIT (Plummer et al. 2004), RVMNT (Plummer et al. 2004), CVMNN (Huaman et al. 2004), SVIET (Nagesha et al. 2003), CVIKT (Nagesha et al. 2003), CVIDT (Lim et al. 2003), CVTNT (Lim et al. 2003)].
The K76T mutation seems to be a very good marker for in vivo CQR (Vathsala et al. 2004) and recently, the SVMNT haplotype has been suggested to show some resistance to Amodiaquine (AQR) (Warhurst 2003).
We performed a cross sectional survey on 88 Laotian samples, collected during a drug efficacy study in the Lao–Myanmar–China border area. The aim of this study was to investigate the distribution of the main Pfcrt-haplotypes in Southeast Asia and in Laos and in particular to investigate the distribution of the SVMNT and CVIEt alleles in an area were AQR was reported in vitro (Yang et al. 1997) and 52%in vivo CQR was reported (author's unpublished data).
Patients and methods
During a drug efficacy study of artemether-lumefantrine and artesunate-mefloquine in treatment of uncomplicated malaria in the north of Laos in 2003 (Stohrer et al. 2004) blood samples were collected on filter paper (Whatmann#3) and stored at −20 °C. The study was conducted in Luang Namtha Province (Laos), a mountainous area located 350 km northwest of Vientiane, bordering Myanmar and China. Each patient (or their guardians) recruited for the in vivo study gave fully informed written consent. Ethical clearance was granted by the Ethics Committee of the Council of Medical Science, Ministry of Health (Vientiane; Laos) and by the Ethics Committee of Humboldt University Berlin (Germany). Parasite DNA was extracted by use of BIO RAD InstaGeneMatrix (BioRad, Germany). DNA-templates were stored at −20 °C until shortly before usage.
Polymerase chain reaction (PCR)
A nested PCR was performed as described in detail elsewhere, to gain a 145 bp fragment (Djimde et al. 2001). The only modifications were that we used a nested TCRD2-primer labelled with biotin on the 5’-end to perform a SSOP-ELISA to distinguish between haplotypes (Alifrangis et al. 2005). PCR conditions for the outer and nested PCR were as follows: 0.2 mM of each dNTPs, 1 μM of the primer set TCRP1/TCRP2, 1.25 units of Ampiqon TEMPase Hot Start DNA polymerase (Bie & Berntsen, Rødovre, Denmark), buffer containing 2.5 mM MgCl2 and 1 μl DNA-Template. The Program used to amplify the DNA-fragments is identical to (Djimde et al. 2001). The reaction was performed using a 96-well format and the reaction mix was covered with one drop of mineral oil. Included in each run were a set of controls consisting of laboratory strains expressing the most common Pfcrt-haplotypes at c72-76: 3D7 (CVMNK), DD2 (CVIET) and 7G8 (SVMNT).
To distinguish between the different Pfcrt-haplotypes we used a PCR based ELISA method recently developed by Alifrangis et al. (2005). This high-throughput method captures the PCR-products on ELISA-plates coated with streptavidin. A following hybridization with sequence specific probes (SSOPs) (Digoxigenin-labelled) detects the Pfcrt-pattern of each sample. Hybridized probes are detected by Anti-DIG-Antibodies and a following substrate reaction using optical density (OPD). OD-Values were read at 490 nm. The OD-values of the samples were scored as follows: infections were considered positive for a certain haplotype when the haplotype OD-value was above the threshold. The OD-values of positive and negative controls varied between experiments, though only rarely compromising specificity. The variation was possibly due to marginal differences in the strength of the probe binding and washing force during the high stringency washes. Thus, no fixed threshold value could be specified and for each haplotype test, a simple analysis of the positive and negative control samples was performed to set a threshold for positivity. For each haplotype analysis, parasite samples were categorized into single, mixed but with one haplotype in majority or mixed infections as follows: infections were considered to be of single genotype when only one SNP was present at OD-values above the threshold of positivity. Samples were considered to be mixed but containing a majority haplotype when the OD-value of the weakly reacting SSOP was less than half the OD-value of the strongly reacting SSOP. Conversely, if the OD-value of the weakly reacting SSOP was higher than half the OD-value of the strongly reacting SSOP the infection was categorized as mixed with no dominant genotype. Samples were scored negative when the OD-value was lower than the threshold. All results were only taken into consideration, when all positive controls were positive and all negative controls were negative.
The Exact binomial 95% Confidence Interval was calculated by EpiTable.
The cross-sectional screening of the Pfcrt-haplotypes was done on P. falciparum slide positive samples from the north of Laos. During an therapeutic efficacy study (artemether-lumefantrine vs. artesunate-mefloquine) from September 2003 to January 2004, 201 patients were tested malaria positive and 108 were recruited for the in vivo study (Stohrer et al. 2004). Within the study group (n = 108) the median age was 10 (2–66) and the sex distribution was 56 female vs. 52 male patients. Of the 108 positive samples 88 were randomly chosen for the cross-sectional survey of which 47.8% reported to have taken other than the study-drugs as self-medication during the current malaria episode (CQ: 35.3%; SP + CQ: 2.3%; Quinine: 10.2%). Eighty-two (93.2%; 95%CI: 85.7–97.5) showed a positive result after the initial PCR and 81 samples were counted positive after the ELISA. Five (6.2%; 95%CI: 2.0–13.8) out of 81 ELISA positive samples indicated mixed infections (SVMNT/CVIET) and were therefore included to the SVMNT and the CVIET group. The distribution of the haplotype-pattern was as shown in Table 1.
Table 1. Frequency of the pfcrt-haplotypes (c72–76) in the north of Laos (n = 87)
|Total|| ||87||100.0|| |
The current drug regime in the Laos is CQ (first-line) and SP (Fansidar, second-line) but a change to artemisinin-based combination therapy (ACT) will follow in the near future (WHO 2005). Our aim was to monitor drug resistance on the molecular level by investigating the Pfcrt-haplotypes and their distribution in an area with reported CQR (in vivo) and AQR (in vitro) (Yang et al. 1997).
One result of our study is that the South American and PNG Pfcrt-haplotype SVMNT is found in Laos and to our knowledge, for the first time on mainland Southeast Asia. It has been proposed that CQR-parasites originated from at least four independent foci: in Asia spreading to Africa, in PNG and two sites in South America, Peru and Colombia (Wootton et al. 2002). Studies showed that the SVMNT haplotype is also the predominant Pfcrt-pattern in India which raised the question, whether independent sweeps due to CQ drug pressure led to similar Pfcrt-haplotypes, in different parts of the world (Vathsala et al. 2004). If this is the case, geographic foci other than the formerly proposed four for the origin of CQR-mutations needs to be considered (Vathsala et al. 2004). India, where the SVMNT-haplotype was predominantly found (Vathsala et al. 2004), could then be one of those foci and from there the SVMNT haplotype could have spread to Laos, and here especially to areas bordering Myanmar which, in this case, might serve as bridge between India and Laos. This migration of different P. falciparum strains could be explained by the common traffic and trading between the bordering countries and Laos and especially between Laos, Myanmar and Thailand within the ‘Golden Triangle’. Studies already showed that P. falciparum strains significantly differ in different regions in Laos, due to the migration of different ethnic minorities (Dittrich et al. 2003). Obviously, the variation of haplotypes could also be due to genetic recombination. Recently it was proposed that SVMNT might be associated with AQR (Warhurst 2003). In vitro resistance (83.5–100%) against Amodiaquine has been reported from the China–Lao and the Myanmar–Lao border area (Yang et al. 1997) and we found a frequency of the SVMNT haplotype of 31%. It is very difficult to compare resistance data optioned by in vitro tests with data gained during in vivo trials, especially with Amodiaquine. Aubouy et al. (2004) showed that in vitro P. falciparum isolates were markedly different in susceptibility to AQ compared to the actual in vivo outcome (5.4%vs. 40.5%). If there is a correlation between AQR and the SVMNT-haplotype as suggested (Warhurst 2003), 31% prevalence of a potential resistant marker is of concern, while AQ is not suitable for the region. Another result of this cross-sectional investigation is that 100% of the isolates showed amino acid change at c76 (K→T) of the pfcrt-gene. This result verifies data earlier obtained from the bordering countries Cambodia (Lim et al. 2003), Myanmar, Viet Nam and Thailand (Pickard et al. 2003), where the prevalence of pfcrt c76-mutations was 88% (Cambodia) and 97% (Thailand, Viet Nam and Myanmar), respectively. Although c76 is widely believed to be a good prediction marker for CQ in vivo resistance, studies also showed that it is not sufficient to forecast the in vivo outcome, possibly due to host factors like acquired immunity (Pillai et al. 2001). Even so, a 100% prevalence of the K76T mutation in Laos has to be of major concern to the local authorities since CQ is the first line drug.
Our findings add one more piece into the jigsaw of Pfcrt-haplotype occurrence, where additional information is clearly needed to investigate the suggested correlation between SVMNT and possibly other polymorphisms of Pfcrt and AQR. New information concerning molecular AQR-markers would be of great value to public health authorities, where reliable data could be used as an additional tool to monitor spreading resistance against malaria drugs.
We thank all participants of the in vivo drug study. Also thanks to Michael Dalgaard and Karen Molbæk for excellent assistance in performing the PCR-ELISA. The author was supported by ‘Gesellschaft für Tropen- und Reisemedizin e.V’ Berlin, Germany.