Previous Article | Next Article ![]()
Antimicrobial Agents and Chemotherapy, November 2008, p. 3883-3888, Vol. 52, No. 11
0066-4804/08/$08.00+0 doi:10.1128/AAC.00431-08
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Liverpool School of Tropical Medicine, Pembroke Place, Liverpool, L35QA,1 Department of Pharmacology and Therapeutics, 70 Pembroke Place, University of Liverpool, Liverpool, L69 3GF, United Kingdom,2 Department of Parasitology, Prince Leopold Institute of Tropical Medicine, Antwerp, Belgium,3 Tropical Diseases Research Center (TDRC), Ndola, Zambia4
Received 1 April 2008/ Returned for modification 26 June 2008/ Accepted 29 July 2008
|
|
|---|
|
|
|---|
The greatest disadvantage of the use of PYR-SDX has been the rapid evolution of drug-resistant parasites. This is mediated by the sequential acquisition of point mutations in the Pfdhfr and Pfdhps genes (8, 25). With respect to PYR, the single Pfdhfr mutation S108N is responsible for low-level resistance and is followed by an N51I or C59R mutation (resulting in a double mutant). A combination of all three mutations gives rise to the highly PYR resistant triple mutant. A fourth point mutation in Pfdhfr, I164L, is found extensively in Southeast Asia and South America, but there is debate over its existence in Africa (5, 11, 12, 17, 29). The presence of the I164L quadruple mutant confers high-level resistance to PYR, and in Africa this mutation would severely compromise the continued use of PYR-SDX. This mutation would also compromise the use of Lapdap, a combination of chlorproguanil (CPG) and dapsone (DDS) (30), and CPG-DDS-artesunate in Africa. The development and use of these compounds have now been halted due to toxicity in children with glucose-6-phosphate dehydrogenase deficiency.
Many studies have looked for the presence of the I164L mutant allele in Africa; Malawi was the first African country to switch from chloroquine to PYR-SDX in 1993. Using conventional PCR, most studies have not detected this mutation (2, 3, 7, 15, 18, 19, 22-24, 26, 27), but it has been reported at low prevalences in five different African countries.
Of particular interest is the study by Alker et al. (1), who used real-time PCR with fluorescent probes specific for the mutation and reported a 4.7% prevalence in parasites collected from human immunodeficiency virus (HIV)-positive pregnant women in Malawi between 2001 and 2003, a finding that they validated more recently using a heteroduplex tracking assay (9). However, on biological grounds, if the quadruple-mutant alleles were present in reasonable proportions, it is hard to imagine that they would not have been selected to high levels by this time. This is the highest prevalence reported, and it is a high priority for the public health that this finding be further evaluated.
The aims of this study were to confirm and validate the sensitivity, specificity, and reproducibility of the assay reported by Alker et al. (1) and to confirm the presence of the mutant 164 allele in parasites collected from the same location in Malawi. We also tested the hypothesis that sustained antifolate use would have resulted in an increased prevalence of the I164L mutant in subsequent years. Finally, we wanted to determine if treatment failure after treatment with PYR-SDX resulted in the selection of this mutation. The prevalence of the I164L allele was also investigated in clinical isolates from Zambia, a neighboring country with a shorter history of PYR-SDX deployment, and in clinical isolates from the Thailand-Myanmar border in Southeast Asia, an area known to have a high prevalence of the I164L mutation (5, 10).
|
|
|---|
Field isolates were obtained from studies conducted in Malawi between 2003 and 2005 (n = 210), in Zambia in 2005 (n = 55), and around the Thailand-Myanmar border in 2005 (n = 50). The Malawian samples were from children less than 5 years old presenting with uncomplicated malaria and were treated as part of a study with PYR-SDX alone or PYR-SDX plus either chloroquine, artesunate, or amodiaquine. The study took place at a government health center 10 km outside of Blantyre, Malawi, where malaria transmission occurs all year round. Parasite isolates were collected prior to treatment and from children who had recurrent parasitemia after treatment. For 57 children, both pre- and posttreatment isolates were available. Genotypic analysis of the msp2 gene showed that approximately one-third of these recurrent parasitemias were due to reinfection and the remainder were due to recrudescence. The pretreatment prevalence of Pfdhfr triple-mutant parasites was 96%, compared to a prevalence of 80% in the parasites described by Alker et al. (1). Details of the Malawian study, which includes Pfdhfr 164 genotyping using a less-sensitive methodology (PCR and allele-specific restriction analysis) have been published elsewhere (4).
The Zambian samples were collected from adults with uncomplicated malaria before treatment as part of a randomized clinical trial with either PYR-SDX or artemether-lumefantrine. The samples were a fair representation of the population, and inclusion/exclusion criteria have been published previously (14, 28). The Thai samples were from adults with P. falciparum malaria before treatment.
Whole blood from the patients in these studies was spotted onto Whatman 3MM filter paper, air dried at room temperature, and stored in individual plastic bags with a desiccant. In addition, for some children in the Malawi study, venous blood was collected in EDTA tubes and stored at –80°C. All the studies contributing samples to this work were conducted under clinical protocols approved by the corresponding institutional review boards.
Isolation and extraction of total DNA. Total genomic DNAs (host and parasite) were extracted either from EDTA-treated whole blood or from blood-spotted filter papers with the QIAamp DNA blood minikit (Qiagen) according to the manufacturer's instructions.
Whole-genome amplification. A high degree of variability in parasitemia and a low recovery rate of parasite genomic DNA from filter paper were observed. Therefore, when the total-parasite DNA concentration was below 10,000 copies per µl, whole-genome amplification by improved primer extension PCR was conducted. This procedure was performed as previously described (6) in order to increase the quantity of DNA and maximize the number of experiments that could be conducted on each sample.
Real-time PCR-based discrimination of Pfdhfr I164L alleles by the method of Alker. In order to assess the sensitivity of the assay, real-time PCR was conducted using the methodology of Alker et al. (1) using a PTC-200 Peltier thermal cycler with a Chromo 4 continuous fluorescence detector (Bio-Rad). This PCR was conducted on reference strains as well as on samples obtained from Malawi that had not undergone whole-genome amplification. The primer and probe sequences are identical to those published previously (1) and were sourced from Applied Biosystems, Cheshire, United Kingdom.
Normalization of Plasmodium DNA by quantification of EF1-
.
Owing to the high degree of variability in parasite DNA content, it was necessary to normalize the samples according to the parasite DNA copy number. This was achieved by quantification of elongation factor 1 alpha (EF1-
) as a marker for the parasite genome copy number. Primers and fluorescent probes specific for the EF1-
gene were designed so as to avoid introns (Table 1).
|
View this table: [in a new window] |
TABLE 1. All primer and probe sequences utilized in this study
|
forward and reverse primers, and 0.25 µM EF1-
probe. The samples were amplified in a PTC-200 Peltier thermal cycler with a Chromo 4 continuous fluorescence detector. The program consisted of an initial activation at 95°C for 10 min, followed by 45 cycles of denaturation at 92°C for 14 s and annealing/extension at 60°C for 60 s.
Amplicons were checked for the presence of nonspecific products by electrophoresis on a 1% agarose gel. Each amplicon was gel purified by using the Promega Gel Wizard prep kit according to the manufacturer's instructions. The copy number of amplicons was then quantified as described previously (20) and diluted to 1, 10, 100, 1,000, 10,000 and 100,000 copies per µl. EF1-
was then quantified alongside clinical samples as described above. Following amplification, the cycle threshold (CT) was determined for each sample and standard, the data plotted, and the equation of the linear regression used to determine the numbers of copies of parasite DNA in the clinical samples. Parasite DNA was then normalized to 100, 250, 500, 1,000, and 10,000 copies per µl for validation of the assay.
Validation of real-time PCR-based discrimination of Pfdhfr I164L alleles. For quantification of the Pfdhfr I164L DNA, mutant (V1/S) and wild-type (K1) genomes were combined in fixed ratios of 100:0, 99:1, 95:5, 90:10, 75:25, 50:50, 25:75, and 0:100. These standards were then amplified, and the CT values for the mutant probe were divided by the corresponding CT values for the wild-type probe. These values were then plotted against log-transformed percentages as a standard curve.
All reactions were carried out in duplicate in a total volume of 25 µl. Each reaction mixture contained 1x Absolute QPCR mix (ABgene), 200 nM each probe, 288 nM forward primer, and 490 nM reverse primer (Table 1). The program consisted of an initial activation at 95°C for 10 min, followed by 45 cycles of denaturation at 92°C for 14 s and annealing/extension at 60°C for 60 s.
Initial experiments were conducted with 100, 250, 500, 1,000, and 10,000 copies per µl of parasite DNA in order to determine the optimum concentration. The intrarun precision of the analyses, expressed as a percentage, was calculated as 100 – [(standard deviation/mean) x 100]. The interrun precision, also expressed as a percentage, was calculated as (calculated log copy number)/(nominal log copy number added) x 100. Determinations were performed using the same amplicon stock solutions. Intrarun precision and accuracy were assessed on six replicates of standards containing 10% and 75% concentrations of the mutant genome. Similarly, interrun precision and accuracy were assessed on six separate runs of standards containing 10% and 75% concentrations of the mutant genome. The limit of detection was defined as the percentage at which the mutant allele could be reliably differentiated from the wild-type allele, and the limit of quantification was defined as values between 90 and 110% for both intra- and interrun accuracy and precision.
For quantification of the percentage of an individual's total parasite population containing the mutant allele, normalized DNAs were amplified as described above. A standard curve was coamplified on each plate, and both standards and samples were assessed at least in duplicate.
|
|
|---|
![]() View larger version (23K): [in a new window] |
FIG. 1. Problems experienced with real-time PCR prior to validation. (A) Real-time PCR trace for the mutant allele with Malawian isolates illustrating the high degree of intersample variability. (B) Real-time PCR trace for wild-type and mutant alleles in an isolate (V1S) known to be entirely mutant for this mutation, illustrating the lack of specificity of the wild-type probe. (C) Real-time PCR trace for wild-type and mutant alleles in an isolate (K1) known to be entirely wild type for this mutation, illustrating the lack of specificity of the mutant probe.
|
(Fig. 2A). A standard curve generated from purified, quantified prerun amplicons was constructed alongside samples (Fig. 2B). Using this methodology, the median copy number per µl of parasite DNA isolated from Malawian samples was 27,776 (range, 4 to 948,664,062). The median level of parasitemia in the Malawian patients was 66,585 parasites per µl (range, 39 to 644,840 parasites per µl). For log-transformed data, a significant correlation was observed between parasitemia and the copy number of isolated parasite DNA (R2 = 0.13; P < 0.0001).
![]() View larger version (16K): [in a new window] |
FIG. 2. Normalization of parasite DNA in order to resolve intersample variability. (A) Representative real-time PCR trace using specific EF1- -directed primers and probes. For the template, EF1- amplicons were gel extracted, quantified, and diluted appropriately prior to the real-time PCR assay, as described in Materials and Methods. (B) Representative standard curve for quantification of EF1- . Samples of DNA from clinical isolates underwent quantification of EF1- , and CT values were read from this curve in order to normalize for the copy number of the parasite genome. (C) Real-time PCR trace for the mutant allele within Malawian isolates following normalization to 10,000 copies of EF1- per reaction. Comparison with Fig. 1A reveals the effective normalization of parasite DNA.
|
![]() View larger version (18K): [in a new window] |
FIG. 3. Quantification of Pfdhfr I164L mutant alleles by real-time PCR. (A) Representative real-time PCR trace for the mutant probe conducted on mixtures of mutant and wild-type alleles. Parasite DNAs from a mutant strain (V1S) and a wild-type strain (K1) were quantified by normalization to EF1- . The DNA of the mutant allele was then diluted appropriately with that of the wild-type allele so as to yield wild-type DNA containing 0% to 100% mutant DNA. (B) Representative real-time PCR trace for the wild-type probe conducted on mixtures of mutant and wild-type alleles (reciprocal of panel A). (C) Representative standard curve used for quantification of the mutant allele in clinical isolates. The CT values obtained from panel A were divided by those obtained from panel B and plotted against the log-transformed percentage of mutant DNA. See the text for the limits of detection, limits of quantification, and inter- and intrarun accuracy and precision.
|
Assessment of Pfdhfr I164L frequency in Malawian, Zambian, and Thai cohorts. In samples where the total-parasite DNA concentration was below 10,000 copies per µl, whole-genome PCR was conducted prior to genotypic analysis. Subsequently, sufficient DNA was obtained from 158 of the 210 Malawian isolates (94 pretreatment isolates and 64 isolates from recurrent parasitemia after treatment), 42 of 55 Zambian isolates, and 38 of 50 Thai isolates. Using the optimized real-time PCR methodology, the frequency of the I164L mutation in these isolates was monitored (Table 2). For the Thai isolates, the I164L mutation was present in 36 out of 38 samples tested. The two remaining samples showed indications that if the mutation was present, then it was below the 5% confidence level of the assay. Conversely, the I164L mutation was below the confidence level in all Malawian and Zambian isolates tested. Furthermore, there was no evidence of the selection of this I164L mutation in any of the 64 Malawian isolates appearing within 42 days of treatment with PYR-SDX (Table 3).
|
View this table: [in a new window] |
TABLE 2. Frequency of the Pfdhfr I164L alleles in isolates from Thailand, Zambia, and Malawi
|
|
View this table: [in a new window] |
TABLE 3. Frequency of different Pfdhfr I164L alleles in Malawian isolates from patients naïve to therapy and from patients having received therapy
|
|
|
|---|
Our approach was found to be both accurate and precise for the detection of the I164L mutation and was able to detect the mutation at a level of 5% of the total parasite population. Furthermore, the assay was able to accurately identify the presence of this mutation in Thai isolates at a frequency comparable to that observed in previous reports (5, 16).
The assay presented here allows improved sensitivity over conventional methodologies, but owing to the incomplete specificity of the probes, it is not possible to quantify the mutation if it is present at less than 5% of the total parasite population within an individual. However, of the Malawian isolates studied here, 64 were parasitological failures after treatment with PYR-SDX, and 57 of these had matched pretreatment isolates (Table 3). Antifolate therapy with PYR-SDX would be expected to select for the I164L mutation if it existed even at low levels in our patient population, yet it was not detected in any of the samples tested. The Malawian samples used in this study were collected from young children who self-presented to a health center situated 10 km away from the hospital where the samples for the Alker study were collected. It is unlikely that the parasites were under different drug pressures. This suggests that the I164L mutation was not preexistent within the samples tested in this study.
Our samples were collected between 2003 and 2005, whereas the samples in the previous study were collected between 2001 and 2003. Therefore, one would expect that in this time an expansion in the mutant population would have occurred under continued selection by high-level PYR-SDX as the first-line treatment in Malawi. Finally, the previous study utilized isolates from pregnant mothers who were HIV positive (1). It is possible that the immunosuppression associated with HIV infection or exposure to cotrimoxazole prophylaxis may have influenced the development of the Pfdhfr I164L mutation or that other, as yet unknown selection pressures were present in this patient group and not in children. It should be emphasized that each sample assayed in this paper could have had less than 5% of the mutant allele, and it would not have been detected. Nonetheless, the rarity of this allele in Africa despite more than a decade of use of PYR-SDX as frontline antimalarial therapy is an intriguing phenomenon, particularly given its rapid selection in Southeast Asia under similar circumstances (5, 16) and the fact that it is easily selected for in in vitro studies (21).
Using a fully validated methodology, we could not identify I164L mutants within clinical isolates from Malawi and Zambia, even in posttreatment failure parasites from Malawi. These data are reassuring, because even though CPG-DDS and CPG-DDS-artesunate have now been withdrawn, PYR-SDX is still used extensively for treatment in Africa and plays a major role in intermittent presumptive therapy programs in pregnancy. Our results are in agreement with the majority of previous reports, and coupled with the need for specialized equipment and the cost associated with real-time PCR, there appears to be no urgent need for field application of this method. The failure of antifolate chemotherapy to select I164L mutant parasites in Africa compared to Southeast Asia is an important phenomenon that clearly requires further investigation of the underlying mechanisms.
This study was funded by support from Gates Malaria Programme Ph.D. studentships to E.O. S.A.W. and P.G.B. are supported by the MRC, BBSRC, and Wellcome Trust. The clinical trial in Malawi was funded by a Wellcome Trust Training Fellowship (066681) awarded to D.J.B. A.O. is supported by the United Kingdom MRC, European Commission, and United Kingdom National Institute of Health Research.
P.A.W. is the chairman, and S.A.W. is an unpaid member, of the MMV development team for the drug "CDA" (Dacart; GSK).
Published ahead of print on 25 August 2008. ![]()
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»