Background

The increasing prevalence of multi-drug resistant parasites threatens to impede efforts to control malaria world-wide [1]. Current in vitro and in vivo methods to monitor the emergence of drug resistance are difficult, costly and labor-intensive. Molecular methods could prove to be useful alternatives [2].

For Plasmodium falciparum, the pfmdr 1 gene product, PGH1, is thought to play an integral role in the mechanism behind parasite resistance to multiple malarial drugs [3]. Both increased gene copy number and single-nucleotide polymorphisms (SNP's) have been associated in epidemiological studies with changes in sensitivity to chloroquine and mefloquine (reviewed in [1, 4, 5]). In two studies from South-east Asia, both increased pfmdr 1 gene copy number and SNPs have been associated with in vitro resistance to mefloquine [6, 7].

In this manuscript, a real-time PCR assay has been developed for the detection of pfmdr 1 SNPs. This assay was developed to be used in conjunction with a previously developed assay for pfmdr 1 gene copy number [6] to assess molecular markers as predictors of mefloquine failure in a clinical study that took place in Sangkhlaburi, Thailand, in 2001–2. The results of the clinical study will be described in a subsequent publication.

Numerous studies have shown discrepancies between in vitro and in vivo tests for antimalarial drug resistance [8–11]. In order to understand why, pfmdr 1 genotypes were determined directly from patient blood and from cultures derived from those patients and compared.

Methods

Patient samples

Subject recruitment. This study received ethical approval from the University of North Carolina IRB and the Thai Ministry of Public Health. Patients presenting with slide-confirmed falciparum malaria to the free Ministry of Public Health malaria clinic or the Kwai River Christian Hospital clinic in Sangklaburi, Thailand from during July 2001 – August 2002 were enrolled. Patients with vivax infections, history of anti-malarial drug use within the past two weeks, bleeding tendency (by self-reported history or based on medical records), or severe/complicated malaria requiring prompt medical management for life support were excluded. In total, 74 patients consented. Blood samples were taken, aliquoted, stored in liquid nitrogen, and transported to the Armed Forces Research Institute for the Medical Sciences (AFRIMS) in Bangkok. At AFRIMS, aliquots were thawed for in vitro culture. 59 of the 74 patient blood samples were successfully cultured as previously described [6, 12]. Parasites were cultured for between 2 and 188 days (mean = 38.5; median = 23). DNA was extracted from patient blood ("pre-culture") and cultured parasites ("post-culture") using the QiaAmp DNA Blood Minikit Blood and Body Fluid Spin Protocol (Qiagen, Valencia, CA) and then shipped on dry ice to the University of North Carolina.

Standard PCR

DNA was PCR-amplified using a MJ Thermocycler (MJ Research, Waltham, MA) as previously described [6]. DNA sequencing was performed at the University of North Carolina Sequencing Core facility using a 3100 Genetic Analyser (Applied Biosystems, Foster City, CA). The sequencing reaction was done using the ABI PRISM™ BigDye™ Terminator Cycle Sequencing Ready Reaction Kit with AmpliTaq® DNA Polymerase, FS (Applied Biosystems).

Real-time PCR

For real-time PCR, pre-culture DNA and post-culture DNA were genotyped using an Applied Biosystems Prism 7000 Sequence Detection System (Applied Biosystems). This assay uses sequence-specific fluorescently-labelled "MGB™ " probes to distinguish SNPs [13–15]. Primers and probes (Table 1) were designed using Primer Express Software, Version 2.0 (Applied Biosystems). Primers were obtained from Qiagen. Fluorescent MGB™ probes were produced by Applied Biosystems and labelled with reporter dyes, 6FAM or VIC, at their 5' ends and non-fluorescent quencher molecules at their 3' ends. Primer and probe concentrations were optimized according to the TaqMan Universal PCR Master Mix technical bulletin (Applied Biosystems) using DNA obtained from P. falciparum strain 3D7 or from patient specimens. 300 nM of both forward and reverse primers were found to be optimal for all SNPs. Optimal probe concentration was determined to be 250 nM for all MGB probes used in this study. Using 10-fold serial dilutions of 3D7 genomic P. falciparum DNA, as few as 8 copies (0.001 ng) of genomic P. falciparum DNA could be detected.

Table 1 Primers and probes used in real-time PCR assay
Full size table

Each real-time PCR amplification reaction contained both the wild type probe labelled with FAM, and the mutant probe labelled with VIC. Presence of either SNP could be imputed from the relative increase of fluorescence of the two fluorophores (Fig. 1). All reactions were done in duplicate.

Figure 1
figure 1

Fluorescent intensity (Rn) of the FAM reporter dye (blue line) compared to VIC reporter dye (green line) as PCR amplification cycles proceed (x axes). Fluorescent intensity is measured as Rn (y-axes), the absolute fluorescence of the reporter dye divided by that of the passive reference dye, Rox. An increase in FAM indicates a wild type population [86(e), 184(f), 1034(g), 1042(h)] while an increase in VIC indicates a mutant population [86(a), 184(b), 1034(c), 1042(d)]. Polyclonal infections may be identified by an increase in fluorescent intensity of both reporter dyes (not shown). Mutant 1034 DNA (c) was obtained from a previous study [6] because none of the samples in this study contained that mutation.

Full size image

Results

Parasite culture

Of the 74 patients enrolled, parasites were successfully cultured from 59. Following treatment, 20 of the 74 admission patients recrudesced. Parasites were successfully cultured from 13 of these 20 recrudescent patients.

Sensitivity and Ssecificity of Real-Time PCR

The sensitivity and specificity of SNP detection using real-time PCR were measured in 22 post-culture DNA samples using DNA sequencing as the gold standard (Table 2). Nineteen cultures were derived from blood of patients upon admission to the study. The other three cultures were obtained from who failed treatment. Two samples contained mixtures of SNPs at single positions based on DNA sequencing. Of the 20 monoclonal samples, 19 were concordant for all pfmdr 1 SNPs when analysed using both methods. The discordant sample was not an artefact since both the real-time PCR and sequencing reactions gave identical results when repeated. The discordant sample contained a single discrepancy at position 184 (phenylalanine by sequencing and tyrosine by real-time PCR). Thus, the sensitivity of real-time PCR method to detect Tyr86 or Asp1042 SNPs was 100% and the sensitivity to detect Phel84 was 92%; there were no Cysl034 mutations detected. Real-time PCR was 100% specific for each mutation. The sensitivity and specificity of the assay to detect all four SNPs were 94% and 100%, respectively.

Table 2 Prevalence in monoclonal culture samples (n = 20) of mutations detected by traditional PCR and sequencing (gold standard) and real-time PCR, and sensitivity and specificity of realtime PCR assay.
Full size table

Two of the 22 samples were found to be mixed infections when sequenced. One sample contained a mixture of Tyrl84 and Phel84. The other sample contained a mixture of Asn86 and His86. The real-time PCR method accurately identified the Tyrl84/Phel84 mixed population. However, because the probes were designed to detect only Asn86 or Tyr86, His86 was not detected by real-time PCR.

Comparison of SNPs pre- and post-culture

In order to determine how the process of culturing parasites in vitro alters genotype, a comparison was made between the pfmdr 1 genotypes of DNA extracted from pre-and post-culture parasites as measured by real-time PCR. Of the paired pre- and post-culture specimens, both members of 68 pairs were successfully amplified by real-time PCR. Fifty-eight of these pairs were obtained from admission patient blood while 10 pairs were derived from recrudescent patients.

Fifty-eight of the pre-culture genotypes were monoclonal, nine pre-culture samples had mixtures of alleles at single positions, and one had mixed SNPs at two positions (184 and 1042). Table 3 shows the number of isolates with specific combinations of pfmdr 1 genotypes pre- and post-culture at each specific locus. Isolates with identical genotype at each locus before and after culture are enumerated in the shaded diagonal. Overall, fifty-four of 68 (79%) pre-culture DNA samples were completely concordant with post-culture DNA at all four polymorphic sites examined. While there was a wide variation in the length of time parasites were cultured before DNA was extracted, there was no correlation between genotype discordances and length of time in culture (data not shown). Five of the 14 discordant isolates exhibited single differences at position 184, and two were different at position 1042. Seven discrepant samples contained mixtures of alleles at one or more sites before and/or after culture. Six samples showed a mixture of alleles pre-culture and single alleles post-culture. This type of selection occurred at amino acid 184 for four isolates, and 1042 for two isolates. Two of these six also had changes at a second position, amino acid 86. Finally, one of the seven samples was monoclonal at aminoacid 184 pre-culture but polyclonal post-culture (and also exhibited a change from Asn to Asp at 1042).

Table 3 Genotype results of DNA from pre-culture and post-culture DNA by real-time PCR
Full size table

Novel His86 mutation identified

In one DNA culture sample neither wild type nor mutant probes revealed amplification during real-time PCR. This post-culture sample was amplified by standard PCR methods and sequenced to genotype the position 86 region. Sequencing this sample revealed a His86.

Discussion

In this paper, a real-time PCR method is described which accurately ascertains parasite pfmdr 1 genotype. This method has high sensitivity (94%) and specificity (100%) for detecting four pfmdr 1 SNPs associated with drug resistance. There was a 21% discordance in the real-time PCR genotype results between DNA samples obtained directly from patient blood and DNA samples obtained from subsequent in vitro cultures.

Real-time PCR offers several advantages over standard PCR methods for genotyping DNA. First, because of the small volume the material, costs for real-time PCR are much lower than for standard PCR and are as little as $0.40 per reaction. Second, using real-time PCR, a single technician can perform and analyse hundreds of reactions per day, thus reducing the labor cost as well. Third, this method reduces the opportunity for post-PCR contamination. Once the sample is prepared with the reagents, amplification and analysis are completed in a closed-tube system. Finally, real-time PCR analysis of genotype is easier and requires less scientific expertise for analysis. Thus, the initial cost of a real-time PCR instrument (~$40,000) would be offset by savings in labor, quality assurance, and materials in labs analysing large numbers of samples.

In this paper, a novel His86 mutation was identified in two samples. To our knowledge, this mutation has never been identified before and may or may not be associated with drug resistance. Further studies on this mutation are needed.

A previous study utilized a novel real-time PCR technology to identify pfmdr 1 SNPs in parasite DNA extracted from clinical blood samples [16]. Although the chemistry of the assay differed from the technique described here, both studies yielded similar results and concluded that real-time PCR may accurately detect Asn86 → Tyr86, Tyr184 → Phe184, Ser1034 → Cys1034 and Asn1042 → Asp1042 in pfmdr 1. In addition, de Monbrison, et al., used this method to genotype Asp1246 → Tyr1246. SNPs at 1246 were not studied here because they have not been found in South-east Asia [6, 7].

Twenty-one percent of samples manifested differences in pfmdr 1 genotype pre- and post-culture. Since most patients are probably infected with a mixture of strains, the genotypes observed post-culture reflects the results of the selection of a subset of strains by in vitro culture conditions. This could manifest itself as loss of strains observed pre-culture or the appearance of strains not observed pre-culture because they were present at levels below detection limits. A much greater change in genotype (74.5%) was observed in a previous study looking at variations in the polymorphic regions of MSP1 and MSP2 (merozoite surface protein 1 and 2) as well as GLURP (glutamate-rich protein) [17], perhaps because these genes are more variable.

In addition to SNP detection, our group has previously described a real-time PCR method to measure pfmdr 1 gene amplification [6]. Thus it is now possible to assess both pfmdr 1 gene copy number and SNPs using real-time PCR making it possible to carry out a complete assessment of pfmdr 1 genetics in large cohorts.

Conclusions

Real-time PCR is a sensitive and specific method to detect pfmdr 1 mutations and gene amplification. Because it is inexpensive and amenable to high-throughput, it could be a useful public health tool.