Skip Navigation

Molecular Markers for Failure of Sulfadoxine-Pyrimethamine and Chlorproguanil-Dapsone Treatment of Plasmodium falciparum Malaria

  1. James G. Kublin1,3,
  2. Fraction K. Dzinjalamala3,4,
  3. Deborah D. Kamwendo3,4,a,
  4. Elissa M. Malkin2,a,
  5. Joseph F. Cortese1,
  6. Lisa M. Martino1,a,
  7. Rabia A. G. Mukadam3,
  8. Stephen J. Rogerson4,a,
  9. Andres G. Lescano2,
  10. Malcolm E. Molyneux4,5,
  11. Peter A. Winstanley5,
  12. Phillips Chimpeni3,4,
  13. Terrie E. Taylor3,6 and
  14. Christopher V. Plowe1
  1. 1Malaria Section, Center for Vaccine Development, University of Maryland School of Medicine
  2. 2Department of International Health, Johns Hopkins School of Hygiene and Public Health, Baltimore
  3. 3Malaria Project
  4. 4Malawi-Liverpool-Wellcome Trust Clinical Research Programme, University of Malawi College of Medicine, Blantyre, Malawi
  5. 5School of Tropical Medicine and Department of Clinical Pharmacology and Therapeutics, University of Liverpool, Liverpool, United Kingdom
  6. 6Department of Medicine, College of Osteopathic Medicine, Michigan State University, East Lansing
  1. aReprints or correspondence: Dr. Christopher V. Plowe, Malaria Section, Center for Vaccine Development, University of Maryland School of Medicine, 685 W. Baltimore St., HSF 480, Baltimore, MD 21201 (cplowe{at}medicine.umaryland.edu).

  • ↵a Present affiliations: Department of Epidemiology, University ofMichigan, Ann Arbor (D.D.K.); Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland (E.M.M.); Harvard Medical School, Boston, Massachusetts (L.M.M.); Department of Medicine, University of Melbourne, Melbourne, Australia (S.J.R.).

 
Next Section

Abstract

Molecular assays for monitoring sulfadoxine-pyrimethamine-resistant Plasmodium falciparum have not been implemented because of the genetic and statistical complexity of the parasite mutations that confer resistance and their relation to treatment outcomes. This study analyzed pretreatment dihydrofolate reductase (DHFR) and dihydropteroate synthase (DHPS) genotypes and treatment outcomes in a double-blind, placebo-controlled trial of sulfadoxine-pyrimethamine and chlorproguanildapsone treatment for uncomplicated P. falciparum malaria. Multiple logistic regression was used to identify mutations that were predictive of treatment failure and to identify interactions and confounding factors. Infections caused by parasites with 3 DHFR mutations and 2 DHPS mutations (the “quintuple mutant”) were associated with sulfadoxine-pyrimethamine treatment failure but not with chlorproguanil-dapsone treatment failure. The presence of a single DHFR mutation (Arg-59) with a single DHPS mutation (Glu-540) accurately predicted the presence of the quintuple mutant. If this model is validated in other populations, it will finally be possible to use molecular markers for surveillance of antifolate-resistant P. falciparum malaria in Africa.

In response to evidence of high rates of chloroquine treatment failure, Malawi revised its national policy for treatment of malaria in 1993, becoming the first African country to adopt the antifolate combination sulfadoxine-pyrimethamine as its first-line antimalarial drug [1]. Several other eastern and southern African countries followed suit in recent years, and sulfadoxine-pyrimethamine is the second-line drug in most of the African countries in which chloroquine is still used.

The combination of the dihydrofolate reductase (DHFR) inhibitor chlorproguanil and the sulfone dapsone (LAPDAP, Glaxo SmithKline) is being considered as an alternative to or replacement for sulfadoxine-pyrimethamine in Africa, because it is effective in patients for whom sulfadoxine-pyrimethamine treatment has failed [2], and, with its short half-life, it exerts less selective pressure for resistant parasites [3, 4]. Health officials charged with making therapy and prophylaxis policies are handicapped by a paucity of current and comprehensive data on resistance and by the limitations of available in vivo and in vitro methods of monitoring resistance [5]. Inexpensive reliablemethods ofmeasuring resistance are needed formonitoring the efficacy of sulfadoxine-pyrimethamine where it is already in use and for predicting its efficacy in areas where it is being considered as a replacement for chloroquine.

Mutations in the genes encoding the target enzymes of the antifolate drugs have been known for several years to be associated with in vitro resistance to these drugs. Mutations in DHFR confer resistance to pyrimethamine and the other DHFR inhibitors, and mutations in dihydropteroate synthase (DHPS) confer resistance to sulfadoxine and the other sulfas and sulfones. Rapid and inexpensive molecular assays for these mutations have been developed, and the many reported associations between these mutations and in vivo resistance have generated recommendations for use of molecular assays to monitor antifolate-resistant Plasmodium falciparum [4–28]. However, none of the burgeoning number of molecular epidemiologic studies has provided specific recommendations that can be used to guide malaria treatment policies.

This lag between advances in our understanding of the genetics and epidemiology of drug resistance and practical application of molecular assays in ways that aid public health is mainly due to the complexity of the results of molecular surveys and the difficulty in relating them to treatment outcomes. With several mutations in 2 genes contributing to in vitro resistance to 2 drugs, it has been difficult to identify a specific set of mutations that predicts the treatment outcomes of antifolate combinations.

For DHFR, a point mutation causing a Ser→Asn change at position 108 confers pyrimethamine resistance in vitro with only a moderate loss of susceptibility to the DHFR inhibitors cycloguanil (the active metabolite of proguanil) and chlorcycloguanil (the active metabolite of chlorproguanil). The addition of the mutations Asn→Ile at codon 51 and/or Cys→Arg at codon 59 results in higher levels of pyrimethamine resistance, again with only modest effects on susceptibility to cycloguanil and chlorcycloguanil. A Ser→Thr mutation at position 108, coupled with an Ala→Val change at position 16, confers resistance to cycloguanil and chlorcycloguanil with only a moderate loss of susceptibility to pyrimethamine. Finally, an Ile→Leu change at position 164 (seen in Southeast Asia and South America in areas of high sulfadoxine-pyrimethamine resistance) combined with Asn-108 and Ile-51 and/or Arg-59 confers high-level resistance to both pyrimethamine and cycloguanil, with a more modest effect on susceptibility to chlorcycloguanil [29–33]. In some areas of South America, a Cys→Arg mutation at position 50 seems to have an effect similar to that of the Arg-59 mutation [15, 34]. In Africa, the DHFR triple mutant Asn-108/Ile-51/Arg-59 has been most strongly associated with resistance to sulfadoxine-pyrimethamine. The other DHFR mutations are extremely rare, or, in the case of the critical Leu-164 mutation, their presence in Africa has not yet been confirmed.

For DHPS, several point mutations are associated with in vitro resistance to sulfadoxine and the other sulfas and sulfones: Ser→Ala at codon 436, Ala→Gly at codon 437, Lys→Glu at codon 540, Ala→Gly at codon 581, and Ser→Phe at codon 436 coupled with either Ala→Thr or Ala→Ser at codon 613 [35, 36]. Among these, Gly-437, followed by Glu-540, is most strongly associated with sulfadoxine-pyrimethamine treatment failure in Africa [4, 15].

The relative importance of mutations in DHFR and DHPS to sulfadoxine-pyrimethamine resistance in vivo has been debated [37, 38]. Parasites with fewer than the 3 DHFR mutations Asn-108, Ile-51, and Arg-59 may be cleared by sulfadoxinepyrimethamine, regardless of DHPS genotype [39]. In the presence of this DHFR triple mutant form, the DHPS genotype increases the likelihood of treatment failure. This model is consistent with field studies that show an association between the prevalence of mutations in both genes and sulfadoxine-pyrimethamine treatment failure rates and that demonstrate selection formutations in both genes after sulfadoxine-pyrimethamine treatment [15, 17, 18, 40]. It is also consistent with a recent study that showed that patients infected with parasites carrying the DHPS double mutant Gly-437/Glu-540 and the DHFR triple mutant had a higher relative risk of treatment failure than did those infected with parasites carrying the DHFR triple mutant alone [4].

The relationship between DHFR and DHPS mutations and antifolate resistance is less complicated in geographic areas in which the genetic complexity of P. falciparum is relatively low, where stable sets of DHFR and DHPS mutations are strongly associated with sulfadoxine-pyrimethamine resistance [17]. In most of sub-Saharan Africa, however, not only do DHFR and DHPS mutations occur in many different combinations, but most infections are polyclonal [26, 41, 42], and, therefore, whether mutations on the 2 genes reside in the same parasites within a single infectioncannotbedetermined.Suchpolyclonalinfections also often are mixed at the loci of interest (i.e., individual infections carry some parasites with mutations at specific DHFR or DHPS codons and some parasites without mutations at those codons).

Because of the complicated nature of the molecular data generated by mutation-specific analysis,most field studies have been limited to reporting frequencies of mutations and, in some cases, describing associations between specific mutations or sets of mutations and different measures of treatment outcome. Univariate analyses have shown that the DHFR triple mutant Asn-108/Ile-51/ Arg-59, the DHPS double mutantGly-437/Glu-540, and the DHFR and DHPS quintuple mutant carrying all 5 of these mutations are all statistically associated with sulfadoxine-pyrimethamine treatment failure in vivo [4]. However, multivariate analyses have not been done to determine which mutations or sets of mutations are independently associated with treatment failure and whether DHFR and DHPS mutations interact as risk factors for antifolate treatment failure.

That intrinsic parasite resistance is not the only factor determining treatment outcome is another obstacle to use of molecular assays for antifolate-resistant malaria. Infections caused by parasites with the same drug-resistance genotypes can have different treatment outcomes [4, 17]. This likely is related to host immunity, as is illustrated by the association between years of exposure to P. falciparum infection and the ability to clear chloroquine-resistant P. falciparum [43]. Other factors, such as subtherapeutic levels of drug, may also contribute to treatment outcomes.

To overcome these obstacles and move molecular assays for DHFR and DHP Sinto the realm of applied tools for surveillance, we analyzed DHFR and DHP Sgenotypes in relation to treatment outcomes in a randomized, double-blind trial of sulfadoxinepyrimethamine and chlorproguanil-dapsone efficacy in treating uncomplicated P. falciparum malaria in Malawian children. We designed a standardized system for interpreting polymerase chain reaction (PCR) results and used multivariate analysis to assess the role of specific sets of DHFR and DHPS mutations in sulfadoxine-pyrimethamine and chlorproguanil-dapsone treatment failure. We then sought to identify a subgroup of mutations that might serve as a simple but reliable marker to predict rates of sulfadoxine-pyrimethamine treatment failure.

Previous SectionNext Section

Methods

Study site and design. Children aged 3 months to 6 years who presented with uncomplicated P. falciparum malaria to the Ndirande Health Centre in Blantyre, Malawi, were enrolled in a prospective, double-blind, placebo-controlled study comparing sulfadoxinepyrimethamine and chlorproguanil-dapsone. The clinical results of this trial will be reported elsewhere (J. Sulo, P. Chimpeni, J. Hatcher, J. G. Kublin, C. V. Plowe, M. E. Molyneux, K. Marsh, T. E. Taylor, W. M. Watkins, and P. A. Winstanley [Sulo et al.], unpublished data). In brief, 500 children were randomly assigned to receive either a single dose of sulfadoxine-pyrimethamine and 2 daily doses of placebo or 3 daily doses of chlorproguanil-dapsone for each clinical malaria episode over the course of 1 year. Treatment outcomes were assessed by passive surveillance, in the form of continuous access to study clinicians for any illnesses, active surveillance by clinical and microscopic examination on posttreatment day 7, and monthly clinical evaluations for the duration of the study. For the purposes of this clinical trial, treatment outcomes were defined as follows: “sensitive”, no illness accompanied by parasitemia on posttreatment days 3–6, no parasites detectable by microscopy at day 7, and no recurrence of parasitemia through day 28, and “treatment failure”, parasitemia with or without illness on day 7. Treatment failures that occurred before day 7 were excluded, because many such early treatment failures in this setting are due to slow clearance of parasites without drug resistance (C.V.P., unpublished data); treatment failures that occurred after day 7 were excluded to avoid the difficulty of distinguishing reinfection from recrudescence in an area in which malaria transmission is ongoing.

Mutation analysis. Mutation-specific nested PCR and/or restriction digestions were used to analyze DHFR and DHPS mutations, as described elsewhere [5, 15, 26]. A detailed description of these methods is available at http://medschool.umaryland.edu/CVD/ plowe.html. In brief, parasite DNA was extracted from dried filterpaper blood samples and analyzed at DHFR codons 51, 59, 108, and 164 and at DHPS codons 437, 540, 581, and 613. Assays were not routinely done for mutations at DHFR codon 16, which are rare in field samples [19], or at DHPS codon 436, because the Ala-436 mutation has not been associated with antifolate resistance in field studies [4, 15] and the Phe-436mutation is rare in field samples [19]. To detect the presence of these rare mutations and to ascertainwhether any new DHFR and DHPS mutations had emerged, 3 posttreatment samples from patients treated with sulfadoxine-pyramethamine and 3 from patients treated with chlorproguanil-dapsone were randomly chosen for DNA sequencing by the Biopolymer Laboratory, University of Maryland School of Medicine.

Codon analysis and interpretation. Codon analysis was done by investigators blinded to clinical outcomes. A novel system for characterizing and coding results of the molecular assays was designed to permit standardized analysis (figure 1). Each DHFR and DHPS codon was characterized as “wild type”, “mixed” (both wildtype and mutant genotype clearly present in the same infection), or “pure” mutant (only mutant genotype detected). As illustrated in figure 1, DHFR genotypes for each infection were then categorized as follows: “wild type”, no mutations detected; “single”, infections involving parasites with a single mutation (Asn-108); “double”, infections involving some parasites with 2 mutations (Asn-108 and Ile-51 and/or Arg-59); “triple mixed”, infections involving some parasites with all 3 mutations; and “triple pure”, infections in which all 3 mutations were detected, at codons 108, 51, and 59, and with no evidence of any parasites with the wild-type sequence at these 3 codons. Similarly, DHPS genotypes were categorized as “wild type”, no mutations detected; “single”, infections involving some parasites with 1 of the DHPS mutations found in Malawi (Gly-437 or Glu-540); “double mixed”, infections involving some parasites with both DHPS mutations; and “double pure”, infections in which both of these mutations were detected and with no evidence of parasites with the wild-type sequence at these codons.

Figure 1
View larger version:
Figure 1

Codon analysis scheme for results of molecular assays for Plasmodium falciparum dihydrofolate reductase (DHFR) and dihydropteroate synthase (DHPS)mutations.As explained in detail in the section “Codon analysis and interpretation” in Methods, results for each infection were coded (e.g., “wild type”, “single”, or “double”) based on theminimum number of mutations whose presence could be inferred on the basis of polymerase chain reaction results and on knowledge of which mutations occur only in the presence of other mutations.

This system for coding molecular results was based on our confirmation that certainmutations occurred only in the presence of specific other mutations. Each infection was coded to designate the minimum set of mutations that could be inferred to be present based on the results of PCR analysis of samples from the patient with that infection. For example, an infection involving parasites that yielded mixed results at DHFR codons 108, 51, and 59 would be coded as a “double mutant”. Because mutations at codons 51 and 59 were only found in the presence of the codon 108 mutation, the presence of the 51 and 59 mutations indicates that some parasites involved in that infection carry at least 2 mutations. However, because it is possible that the infection included some parasites with mutations at codons 108 and 51 and other parasites with mutations at 108 and 59 but no parasites with mutations at all 3 codons, the definite presence of the DHFR triple mutant cannot be inferred. Similarly, an infection with mixed results at DHPS codons 437 and 540 was coded as “single mutant”, because both of these mutations can occur independently, whereas an infection with only mutant results at codon 437 and mixed results at codon 540 was coded as a “DHPS mixed double mutant”, because all parasites had the 437 mutation and some also had the 540 mutation. Finally, combined DHFR and DHPS genotypes were categorized as “single”, “double”, “triple”, “quadruple”, “quintuplemixed”, or “quintuple pure”, as shown in table 1. For example, DHFR and DHPS double mutant genotypes include parasites with DHFR double mutant and DHPS wild-type, DHFR single mutant and DHPS single mutant, and DHFR wild-type and DHPS double mutant genotypes.

To measure the association between DHFR and DHPS genotypes and treatment outcomes, the genotypes of parasites causing 49 clinical episodes that resulted in treatment failure were compared with those of the parasites causing 118 randomly selected infections that resulted in sensitive outcomes. Testing was done on samples obtained before treatment. The relationships between genotypes and outcomes were analyzed using a case-control design. All patients with day-7 treatment failures were included in the analysis as case patients and were compared with randomly selected control subjects with sensitive outcomes. The number of control subjects chosen was approximately twice the number of case patients, because statistical power in case-control studies improves significantly when the ratio of control subjects to case patients is increased from 1:1 to 2:1 [44]. To develop a simplified model for using molecular markers for surveillance of sulfadoxine-pyrimethamine resistance, we assessed the sensitivity and specificity of use of the presence of single and paired DHFR and DHPS mutations to predict the presence of the DHFR and DHPS pure quintuple mutant in 149 samples (among the 167 selected for analysis) for which all 5 codons of interest were successfully characterized. To determine the baseline prevalence in this population of the mutations most predictive of treatment failure, we measured the frequency of DHFR Ile-51 and Arg-59 and DHPS Glu-540 in pretreatment samples from 100 patients randomly selected from the whole study population without regard for treatment outcome.

Statistical analysis. We used χ2 tests to assess the significance of the association between mutations and resistance in the univariate analysis. Fisher's exact test was used to correct the χ2 tests when needed. Strength of association was evaluated using odds ratios (ORs). Multiple logistic regression models were fitted to assess the independent effects of the DHPS and DHFR mutations and to allow adjustment for hemoglobin levels. Likelihood ratio tests were used to evaluate the significance of each covariate in the model, and Wald tests were used to assess significance when covariates had > 2 levels. The validity of mutations in 1 or 2 codons as predictors of quintuple mutations was assessed by calculation of sensitivity and specificity. Pure or mixed mutations and pure-only mutations were independently evaluated. Validity was considered to be optimal when the sum of sensitivity and specificity was maximized. All analyses were done using Stata 7.0 and S-Plus 4.5 (Math- Soft) software. All confidence levels were set at 95%.

Previous SectionNext Section

Results

We successfully analyzed 84 pretreatment samples of blood from patients with P. falciparum infections who were treated with sulfadoxine-pyrimethamine and 65 samples from patients who were treated with chlorproguanil-dapsone for DHFR mutations at all 3 of codons 108, 51, and 59 and DHPS mutations at codons 437 and 540. In the clinical study, the rate of failure for sulfadoxine-pyrimethamine treatment was higher (20.5%) than that for chlorproguanil-dapsone treatment (5.1%) (P > :01), and subjects were equally compliant with the sulfadoxine-pyrimethamine regimen and the chlorproguanil-dapsone regimen (Sulo et al., unpublished data). Fewer samples from patients treated with chlorproguanil-dapsone were analyzed, because the number of treatment failures in this group was low. Figure 2 shows that themajority of treatments succeeded when <3 DHFR mutations, <2 DHPS mutations, or <5 mutations total were detected. Once these thresholds were exceeded, the majority of patients receiving sulfadoxine-pyrimethamine experienced treatment failure.

The DHFR triple mixed and triple pure mutant genotypes were associated with sulfadoxine-pyrimethamine treatment failure but not with chlorproguanil-dapsone treatment failure (figure 2 and tables 2 and 3). The DHPS double mutant Gly-437/Glu-540, both in mixed and in pure form, was also associated with sulfadoxinepyrimethamine but not with chlorproguanil-dapsone treatment failure. The parasite genotypemost highlymutated in both DHFR and DHPS in this setting has mutations in DHFR at codons 108, 51, and 59 and in DHPS at codons 437 and 540. This quintuple mutant, in either pure or mixed form, was the genotype most strongly associated with sulfadoxine-pyrimethamine treatment failure; the OR associated with this mutant was nearly twice that associated with the DHFR triple mutant in mixed or pure form and nearly 4 times that associated with the DHPS double mutant in mixed or pure form (table 2). The risk of treatment failure when the parasite causing the initial infection carried the quintuple mutant was markedly higher than the nonsignificant risk of failure when the parasite had the DHFR triple mutant alone (OR, 1.69; P = :549) or the DHPS double mutant alone (OR, 0.41; P = :449).Astrong statistical interaction was observed between the DHFR triple mutant and the DHPS double mutant as risk factors for treatment failure (OR, 18.9; P = :037). The quintuple mutant was not associated with chlorproguanil-dapsone treatment failure.

Figure 2
View larger version:
Figure 2

Dihydrofolate reductase (DHFR) and dihydropteroate synthase (DHPS) genotypes and clinical outcomes of treatment for uncomplicated Plasmodium falciparum malaria with sulfadoxine-pyrimethamine (SP) and chlorproguanil-dapsone (CD). Black columns, treatment failure; gray columns, sensitive to treatment. 1, Single mutant; 2, double mutant; 3, triple mutant; 4, quadruple mutant; 5, quintuple mutant; M, mixed; P, pure; W, wild type.

Table 1
View larger version:
Table 1

Classification system for dihydropteroate synthase (DHPS) and dihydrofolate reductase (DHFR) genotypes in Plasmodium falciparum.

Table 4
View larger version:
Table 4

Risk of developing malaria caused by sulfadoxine-pyrimethamine- resistant or -sensitive Plasmodium falciparum, by dihydropteroate synthase (DHPS) and dihydrofolate reductase (DHFR) mutations.

Table 3
View larger version:
Table 3

Risk of developing malaria caused by chlorproguanil-dapsone-resistant or -sensitive Plasmodium falciparum, by dihydropteroate synthase (DHPS) and dihydrofolate reductase (DHFR) mutations.

The role of factors other than DHFR and DHPS genotype in sulfadoxine-pyrimethamine treatment failure was assessed by adjusting the risk of treatment failure related to DHFR and DHPS mutations for hematocrit, age, and parasite density at the time of treatment. Lower hematocrit was significantly associated with risk of sulfadoxine-pyrimethamine treatment failure (table 4; multivariate adjusted OR, 4.0; P = :04), but neither age nor parasite density at the time of infection was associated with treatment failure (data not shown).

Table 4
View larger version:
Table 4

Risk of developing malaria caused by sulfadoxine-pyrimethamine-resistant or -sensitive Plasmodium falciparum, by dihydropteroate synthase (DHPS) and dihydrofolate reductase (DHFR) mutations, adjusted by baseline hematocrit.

Some infections appeared to be caused by parasites with genotypes of mixed form at codon 108 but mutant form at codons 51 and 59. Because direct DNA sequencing has not confirmed the existence of wild-type Ser-108 combined with mutant Ile- 51 and/or Arg-59, it is likely that these results are the result of different PCR detection thresholds for the different mutations. No mutations at DHFR position 164 or DHPS position 581 or 613 were detected, and no new mutations were found in the samples that were sequenced.

We measured the sensitivity and specificity of the presence of each individualmutation aswell as of all possible pairs of mutations for predicting the presence of the quintuplemutant among 149 infections analyzed at all 5 codons (table 5). When present in either mixed or pure mutant form, all mutations and pairs of mutations were 100% sensitive as surrogate markers for the quintuple mutant, although specificities were poor. Specificities were markedly improved at a moderate cost to sensitivity by the use of pure mutant forms as predictors of the quintuple mutant. The combination of DHFR Arg-59 and DHPS Glu-540 in pure form was the most valid predictor of the quintuple mutant, with 92% sensitivity and 90% specificity (table 5).

Table 5
View larger version:
Table 5

Sensitivity and specificity of single and double mutations as predictors of quintuple mutations in malaria-causing Plasmodium falciparum.

We randomly selected 100 pretreatment samples to measure the baseline prevalence of the DHFR and DHPSmutations in this study population. On the basis of their correlation with the presence of the quintuple mutant, the DHFR Ile-51/Arg-59 and the DHPS Glu-540 mutations were analyzed. In this randomly selected group of samples, the combination of DHFR Ile-51 and DHPS Glu-540 was present in 70% of the samples, of DHFR Arg-59 and DHPS Glu-540 in 65%, and of all 3 mutations in 64%. The DHFR triple mutant (indicated by the presence of DHFRIle-51/Arg-59), at 78%, was more common than the quintuple mutant (indicated by the presence of DHFR Ile-51/Arg-59 and DHPS Glu-540), at 64%. Among the 50 samples from patients who were treated with sulfadoxine-pyrimethamine in this randomly selected subset, 30% were from patients who experienced treatment failure.

Previous SectionNext Section

Discussion

The presence of 3 mutations in DHFR and 2 mutations in DHPS in P. falciparum was strongly associated with failure of sulfadoxine-pyrimethamine but not chlorproguanil-dapsone treatment. Both the DHFR triple mutant Asn-108/Ile-51/Arg-59 and the DHPS double mutant Gly-437/Glu-540 were independently associated with sulfadoxine-pyrimethamine failure, and logistic regression showed that the 2 sets of mutations interact and that the presence of all 5 mutations (the quintuple mutant) was most strongly associated with treatment failure. These results are consistent with the idea that the DHFR triple mutant is necessary but not sufficient for sulfadoxine-pyrimethamine treatment failure but that DHPS mutations and possibly other factors determine whether treatment failure occurs in the presence of the triplemutant [39]. Although the DHFR triple mutant is strongly associated with sulfadoxine-pyrimethamine treatment failure, the quintuple mutant is less prevalent and more specific, making it a more suitable marker for treatment failure. BecauseDHFRandDHPSmutations occur in a stepwise manner, assays for carefully chosen single DHFR and DHPS mutations can predict the presence of the quintuple mutant. This would eliminate the need to perform assays for more than 2 mutations to assess levels of sulfadoxinepyrimethamine treatment failure. With a new, standardized system for classifying results of molecular assays, we now have a model that, if its use in other populations is validated, can be applied directly to surveillance of sulfadoxine-pyrimethamine treatment failure.

Our findings confirm that the P. falciparum mutations associated with sulfadoxine-pyrimethamine resistance in Africa do not impair the clinical efficacy of chlorproguanil-dapsone, a recently developed antifolate combination shortly to be subjected to regulatory review. Chlorproguanil-dapsone is likely to be introduced in areas where sulfadoxine-pyrimethamine resistance has reached unacceptable levels. The low rates of chlorproguanil-dapsone treatment failure in the face of a prevalence of the quintuple mutant of >60% supports the view that this antifolate combination will remain effective unless and until the DHFR mutation Leu- 164 appears in Africa. This mutation is associated with highlevel resistance to the active metabolite of chlorproguanil in vitro and with high rates of sulfadoxine-pyrimethamine treatment failure in South America and Southeast Asia [15, 17, 34]. Continued surveillance for this mutation in Africa is warranted, particularly in areas with high rates of sulfadoxine-pyrimethamine treatment failure. It would also be prudent to continue surveillance for the DHPS mutation at codon 581, which is associated with high rates of sulfadoxine-pyrimethamine resistance in South America and SoutheastAsia but remains rare in Africa [15, 17].

Where resources for molecular assays are limited, we recommend limiting surveillance for these rare mutations to periodic surveys at sentinel sites or where significant chlorproguanildapsone resistance is observed. Surveillance for other mutations that are rarely found in nature or do not appear to be associated with resistance in epidemiologic surveys, including DHFR Val- 16 and Thr-108 and DHPS mutations at codons 436 and 613, can likewise be limited to settings in which there is a reason to suspect their presence. At sites that have the capability to do DNA sequencing, periodic sequencing of subsets of samples to monitor for the appearance of rare or new mutations is recommended.

How, then, should molecular assays be used for surveillance of antifolate-resistant P. falciparum and to inform treatment and prevention policy? We recommend that regional research centers in Africa, several of which already can perform these assays, start with surveys for all 5 of the DHFR and DHPS mutations commonly found in Africa. In this study, we found that DHFR Arg-59 and DHPS Glu-540 in pure form can be used to predict the presence of the quintuple mutant with a high degree of accuracy. Because the ability to distinguish mixed from pure infections can vary with different concentrations of template DNA and under different PCR conditions, and because the relative prevalence of these 2 mutations varies in different settings, the choice of surrogate markers for the quintuple mutant needs to be validated in other countries and laboratories. It then will be necessary to relate the prevalence of the molecular markers to the prevalence of treatment failure.

Amodel thatusesamolecularmarker for chloroquineresistance was recently described [45].Dividing the prevalence of themarker (in the case of sulfadoxine-pyrimethamine resistance, DHFR Arg-59 accompanied by DHPS Glu-540 in pure form) by the prevalence of parasitologic resistance yields a genotype-resistance index (GRI). Similarly, a genotype-failure index (GFI) can be calculated to predict rates of therapeutic efficacy by using newer definitions recommended by theWorld Health Organization (WHO) [46]. These indices must be adjusted for age by logistic regression analysis before they will be comparable for studies done at different times and at sites with different rates of malaria transmission and host immunity [45]. Once established, these indices may permit prediction of rates of resistance to sulfadoxine-pyrimethamine and treatment failure, based on crosssectional surveys for just 2 mutations.

To permit this first evaluation of our model with aminimum of confounding, we used an unusual definition of treatment failure to allow the highest degree of confidence that treatment failures were due to drug resistance and not to slow parasite clearance or reinfection. To establish and validate GRIs and GFIs for sulfadoxine- pyrimethamine, it will be necessary to use standardized definitions of parasitologic resistance and treatment failure. Most in vivo malaria drug efficacy studies use variations of pr tocols recommended by the WHO. These protocols have limitations, including overestimation of high-level parasitologic resistance and early treatment failure, and may need to be revised [47]. Universal application of these methods will require a harmonization of measures of drug treatment outcome.Nevertheless, by using the definition of treatment failure adopted for this study, a preliminary estimation of GFI can still be made: Among the 100 randomly selected samples that were assayed for DHFR and DHPS mutations, 50 were from patients who had been treated with sulfadoxine-pyrimethamine, and, in 30% of those, treatment failed. The GFI for DHFR Arg-59 and DHPS Glu-540 is the prevalence of this combination of mutations divided by the prevalence of failure, or 65% divided by 30%, yielding a GFI of 2.2. If similar GFIs are validated and are stable after age adjustment, they could be used to predict sulfadoxine treatment failure rates of approximately one-half of the prevalence of the DHFR Arg-59 and DHPS Glu-540 genotype in a given setting.

This study confirms thatmutations in P. falciparum DHPS and DHFR are important predictors of sulfadoxine-pyrimethamine resistance in Malawi, the first African country to adopt this drug combination as its front-line treatment against malaria. Further research on and discussion of the relative importance of these mutations to resistance and the exact mechanisms by which they exert their effect continue to be of scientific interest but will not speed the process of developing practical tools for monitoring antifolate resistance in endemic areas. We recommend multisite collaboration and use of harmonized clinical and molecular protocols to validate and refine the model proposed here. Techniques that offer practical advantages to in vivo studies and that can be applied meaningfully in the field will only become available if this type of simplified model for using molecular assays to conduct surveillance for drug-resistant P. falciparum can be validated and implemented.

Previous SectionNext Section

Acknowledgements

We thank Ernest Tompson, Eunice Matemba, and Esther Mwagomba, for assistance with this study.

Previous SectionNext Section

Footnotes

  • Presented in part: American Society of Tropical Medicine and Hygiene 48th annual meeting, Washington, DC, December 1999 (abstract 454).

  • Informed consent was obtained from subjects or their parents or guardians in accordance with the human experimentation guidelines of the US Department of Health and Human Services and of the authors' institutions.

  • Financial support: National Institutes of Health (grants AI-40539 and AI-44824); Department of Medicine, University of Maryland School of Medicine; Wellcome Trust, UK (fellowships to S.J.R. and M.E.M.); clinical trial funding from the World Health Organization (TDR 910151) in collaboration with SmithKline Beecham Pharmaceuticals.

  • Received June 1, 2001.
  • Revision received September 17, 2001.
Previous Section
 

References

  1. 1.↵
    1. Bloland PB,
    2. Lackritz EM,
    3. Kazembe PN,
    4. Were JB,
    5. Steketee R,
    6. Campbell CC
    . Beyond chloroquine: implications of drug resistance for evaluating malaria therapy efficacy and treatment policy in Africa. J Infect Dis 1993;167:932-7.
    Abstract/FREE Full Text
  2. 2.↵
    1. Mutabingwa T,
    2. Nzila AM,
    3. Mberu EK,
    4. et al
    . Chlorproguanil-dapsone for treatment of drug resistant falciparum malaria in Tanzania. Lancet 2001;358:1218-23.
    CrossRefMedlineWeb of Science
  3. 3.↵
    1. Watkins WM,
    2. Mosobo M
    . Treatment of Plasmodium falciparum malaria with pyrimethamine-sulfadoxine: selective pressure for resistance is a function of long elimination half-life. Trans R Soc Trop Med Hyg 1993;87:75-8.
    CrossRefMedlineWeb of Science
  4. 4.↵
    1. Nzila AM,
    2. Mberu EK,
    3. Sulo J,
    4. et al
    . Towards an understanding of the mechanism of pyrimethamine-sulfadoxine resistance in Plasmodium falciparum: genotyping of dihydrofolate reductase and dihydropteroate synthase of Kenyan parasites. Antimicrob Agents Chemother 2000;44:991-6.
    Abstract/FREE Full Text
  5. 5.↵
    1. Plowe CV,
    2. Djimde A,
    3. Bouare M,
    4. Doumbo O,
    5. Wellems TE
    . Pyrimethamine and proguanil resistance-conferring mutations in Plasmodium falciparum dihydrofolate reductase: polymerase chain reaction methods for surveillance in Africa. Am J Trop Med Hyg 1995;52:565-8.
    Abstract/FREE Full Text
  6. 6.
    1. Gyang FN,
    2. Peterson DS,
    3. Wellems TE
    . Plasmodium falciparum: rapid detection of dihydrofolate reductase mutations that confer resistance to cycloguanil and pyrimethamine. Exp Parasitol 1992;74:470-2.
    CrossRefMedlineWeb of Science
  7. 7.
    1. Peterson AS,
    2. Di Santi SM,
    3. Povoa M,
    4. Calvosa VS,
    5. Do Rosario VE,
    6. Wellems TE
    . Prevalence of the dihydrofolate reductase Asn-108 mutation as the basis for pyrimethamine-resistant falciparum malaria in the Brazilian Amazon. Am J Trop Med Hyg 1991;45:492-7.
    Abstract/FREE Full Text
  8. 8.
    1. Wang P,
    2. Brooks DR,
    3. Sims PF,
    4. Hyde JE
    . Amutation-specific PCR system to detect sequence variation in the dihydropteroate synthetase gene of Plasmodium falciparum. Mol Biochem Parasitol 1995;71:115-25.
    CrossRefMedlineWeb of Science
  9. 9.
    1. Eldin DP,
    2. Abdallah B,
    3. Dje MK,
    4. Basco LK,
    5. Le Bras J,
    6. Mazabraud A
    . Use of a semi-nested PCR diagnosis test to evaluate antifolate resistance of Plasmodium falciparum isolates. Mol Cell Probes 1995;9:391-7.
    CrossRefMedlineWeb of Science
  10. 10.
    1. Plowe CV,
    2. Djimde A,
    3. Wellems TE,
    4. Kouriba B,
    5. Doumbo OK
    . Community pyrimethamine use and prevalence of resistant Plasmodium falciparum genotypes in Mali: a model for deterring resistance. Am J Trop Med Hyg 1996;55:467-71.
    Abstract/FREE Full Text
  11. 11.
    1. Curtis J,
    2. Duraisingh MT,
    3. Trigg JK,
    4. Mbwana H,
    5. Warhurst DC,
    6. Curtis CF
    . Direct evidence that asparagine at position 108 of the Plasmodium falciparum dihydrofolate reductase is involved in resistance to antifolate drugs in Tanzania. Trans R Soc Trop Med Hyg 1996;90:678-80.
    CrossRefMedlineWeb of Science
  12. 12.
    1. Zindrou S,
    2. Dung NP,
    3. Sy ND,
    4. Skold O,
    5. Swedberg G
    . Plasmodium falciparum: mutation pattern in the dihydrofolate reductase-thymidylate synthase genes of Vietnamese isolates, a novel mutation, and coexistence of two clones in a Thai patient. Exp Parasitol 1996;84:56-64.
    CrossRefMedlineWeb of Science
  13. 13.
    1. Wang P,
    2. Lee CS,
    3. Bayoumi R,
    4. Djimde A,
    5. et al
    . Resistance to antifolates in Plasmodium falciparum monitored by sequence analysis of dihydropteroate synthase and dihydrofolate reductase alleles in a large number of field samples of diverse origins. Mol Biochem Parasitol 1997;89:161-77.
    CrossRefMedlineWeb of Science
  14. 14.
    1. Edoh D,
    2. Mshinda H,
    3. Jenkins J,
    4. Burger M
    . Pyrimethamine-resistant Plasmodium falciparum parasites among Tanzanian children: a facility-based study using the polymerase chain reaction. Am J Trop Med Hyg 1997;57:342-7.
    Abstract/FREE Full Text
  15. 15.↵
    1. Plowe CV,
    2. Cortese JF,
    3. Djimde A,
    4. et al
    . Mutations in Plasmodium falciparum dihydrofolate reductase and dihydropteroate synthase and epidemiologic patterns of pyrimethamine-sulfadoxine use and resistance. J Infect Dis 1997;176:1590-6.
    Abstract/FREE Full Text
  16. 16.
    1. Jelinek T,
    2. Ronn AM,
    3. Curtis J,
    4. et al
    . High prevalence of mutations in the dihydrofolate reductase gene of Plasmodium falciparum in isolates from Tanzania without evidence of an association to clinical sulfadoxine/ pyrimethamine resistance. Trop Med Int Health 1997;2:1075-9.
    CrossRefMedlineWeb of Science
  17. 17.↵
    1. Kublin JG,
    2. Witzig RS,
    3. Shankar AH,
    4. et al
    . Molecular assays for surveillance of antifolate-resistant malaria. Lancet 1998;351:1629-30.
    CrossRefMedlineWeb of Science
  18. 18.↵
    1. Curtis J,
    2. Duraisingh MT,
    3. Warhurst DC
    . In vivo selection for a specific genotype of dihydropteroate synthetase of Plasmodium falciparum by pyrimethamine-sulfadoxine but not chlorproguanil-dapsone treatment. J Infect Dis 1998;177:1429-33.
    Abstract/FREE Full Text
  19. 19.↵
    1. Plowe CV,
    2. Kublin JG,
    3. Doumbo OK
    . P. falciparum dihydrofolate reductase and dihydropteroate synthase mutations: epidemiology and role in clinical resistance to antifolates. Drug Resist Updat 1998;1:389-96.
    CrossRefMedlineWeb of Science
  20. 20.
    1. Shaio MF,
    2. Wang P,
    3. Lee CS,
    4. Sims PF,
    5. Hyde JE
    . Development and comparison of quantitative assays for the dihydropteroate synthetase codon 540 mutation associated with sulfadoxine resistance in Plasmodium falciparum. Parasitology 1998;116:203-10.
  21. 21.
    1. Duraisingh MT,
    2. Curtis J,
    3. Warhurst DC
    . Plasmodium falciparum: detection of polymorphisms in the dihydrofolate reductase and dihydropteroate synthetase genes by PCR and restriction digestion. Exp Parasitol 1998;89:1-8.
    CrossRefMedlineWeb of Science
  22. 22.
    1. Basco LK,
    2. Ringwald P
    . Molecular epidemiology of malaria in Yaounde, Cameroon. II. Baseline frequency of point mutations in the dihydropteroate synthase gene of Plasmodium falciparum. Am J Trop Med Hyg 1998;58:374-7.
    Abstract
  23. 23.
    1. Basco LK,
    2. Ringwald P
    . Molecular epidemiology of malaria in Yaounde, Cameroon. V. Analysis of the omega repetitive region of the Plasmodium falciparum cg2 gene and chloroquine resistance. Am J Trop Med Hyg 1999;61:807-13.
    Abstract
  24. 24.
    1. Diourte Y,
    2. Djimde A,
    3. Doumbo OK
    . Pyrimethamine-sulfadoxine efficacy and selection for mutations in Plasmodium falciparum dihydrofolate reductase and dihydropteroate synthase in Mali. Am J Trop Med Hyg 1999;60:475-8.
    Abstract
  25. 25.
    1. Urdaneta L,
    2. Plowe C,
    3. Goldman I,
    4. Lal AA
    . Point mutations in dihydrofolate reductase and dihydropteroate synthase genes of Plasmodium falciparum isolates from Venezuela. Am J Trop Med Hyg 1999;61:457-62.
    Abstract
  26. 26.↵
    1. Doumbo OK,
    2. Kayentao K,
    3. Djimde A,
    4. et al
    . Rapid selection of Plasmodium falciparum dihydrofolate reductase mutants by pyrimethamine prophylaxis. J Infect Dis 2000;182:993-6.
    Abstract/FREE Full Text
  27. 27.
    1. Vasconcelos KF,
    2. Plowe CV,
    3. Fontes CJ,
    4. et al
    . Mutations in Plasmodium falciparum dihydrofolate reductase and dihydropteroate synthase of isolates from the Amazon region of Brazil. Mem Inst Oswaldo Cruz 2000;95:721-8.
    MedlineWeb of Science
  28. 28.
    1. Biswas S,
    2. Escalante A,
    3. Chaiyaroj S,
    4. Angkasekwinai P,
    5. Lal AA
    . Prevalence of point mutations in the dihydrofolate reductase and dihydropteroate synthetase genes of Plasmodium falciparum isolates from India and Thailand: a molecular epidemiologic study. Trop Med Int Health 2000;5:737-43.
    CrossRefMedlineWeb of Science
  29. 29.↵
    1. Peterson DS,
    2. Walliker D,
    3. Wellems TE
    . Evidence that a point mutation in dihydrofolate reductase-thymidylate synthase confers resistance to pyrimethamine in falciparum malaria. Proc Natl Acad Sci USA 1988;85:9114-8.
    Abstract/FREE Full Text
  30. 30.
    1. Cowman AF,
    2. Morry MJ,
    3. Biggs BA,
    4. Cross GA,
    5. Foote SJ
    . Amino acid changes linked to pyrimethamine resistance in the dihydrofolate reductase-thymidylate synthase gene of Plasmodium falciparum. Proc Natl Acad Sci USA 1988;85:9109-13.
    Abstract/FREE Full Text
  31. 31.
    1. Zolg JW,
    2. Plitt JR,
    3. Chen GX,
    4. Palmer S
    . Point mutations in the dihydrofolate reductase-thymidylate synthase gene as the molecular basis for pyrimethamine resistance in Plasmodium falciparum. Mol Biochem Parasitol 1989;36:253-62.
    CrossRefMedlineWeb of Science
  32. 32.
    1. Peterson DS,
    2. Milhous WK,
    3. Wellems TE
    . Molecular basis of differential resistance to cycloguanil and pyrimethamine in Plasmodium falciparum malaria. Proc Natl Acad Sci USA 1990;87:3018-22.
    Abstract/FREE Full Text
  33. 33.↵
    1. Foote SJ,
    2. Galatis D,
    3. Cowman AF
    . Amino acids in the dihydrofolate reductase-thymidylate synthase gene of Plasmodium falciparum involved in cycloguanil resistance differ from those involved in pyrimethamine resistance. Proc Natl Acad Sci USA 1990;87:3014-7.
    Abstract/FREE Full Text
  34. 34.↵
    1. Cortese JF,
    2. Plowe CV
    . Antifolate resistance due to new and known Plasmodium falciparum dihydrofolate reductase mutants expressed in yeast. Mol Biochem Parasitol 1998;94:205-14.
    CrossRefMedlineWeb of Science
  35. 35.↵
    1. Brooks DR,
    2. Wang P,
    3. Read M,
    4. Watkins WM,
    5. Sims PF,
    6. Hyde JE
    . Sequence variation of the hydroxymethyldihydropterin pyrophosphokinase: dihydropteroate synthase gene in lines of the human malaria parasite, Plasmodium falciparum, with differing resistance to sulfadoxine. Eur J Biochem 1994;224:397-405.
    MedlineWeb of Science
  36. 36.↵
    1. Wang P,
    2. Read M,
    3. Sims PF,
    4. Hyde JE
    . Sulfadoxine resistance in the human malaria parasite Plasmodium falciparum is determined by mutations in dihydropteroate synthetase and an additional factor associated with folate utilization. Mol Microbiol 1997;23:979-86.
    CrossRefMedlineWeb of Science
  37. 37.↵
    1. Sims P,
    2. Wang P,
    3. Hyde JE
    . On “The efficacy of antifolate antimalarial combinations in Africa”. Parasitol Today 1998;14:136-7.
    CrossRefMedlineWeb of Science
  38. 38.↵
    1. Watkins WM,
    2. Mberu EK,
    3. Winstanley PA,
    4. Plowe CV
    . More on “The efficacy of antifolate antimalarial combinations in Africa”. Parasitol Today 1999;15:131-2.
    CrossRefMedlineWeb of Science
  39. 39.↵
    1. Wang P,
    2. Brobey RK,
    3. Horii T,
    4. Sims PF,
    5. Hyde JE
    . Utilization of exogenous folate in the human malaria parasite Plasmodium falciparum and its critical role in antifolate drug synergy. Mol Microbiol 1999;32:1254-62.
    CrossRefMedlineWeb of Science
  40. 40.↵
    1. Wang P,
    2. Lee CS,
    3. Bayoumi R
    . Resistance to antifolates in Plasmodium falciparum monitored by sequence analysis of dihydropteroate synthetase and dihydrofolate reductase alleles in a large number of field samples of diverse origins. Mol Biochem Parasitol 1997;89:161-77.
    CrossRefMedlineWeb of Science
  41. 41.
    1. Gupta S,
    2. Trenholme K,
    3. Anderson RM,
    4. Day KP
    . Antigenic diversity and the transmission dynamics of Plasmodium falciparum. Science 1994;263:961-3.
    Abstract/FREE Full Text
  42. 42.↵
    1. Farnert A,
    2. Rooth I,
    3. Svensson Å,
    4. Snounou G,
    5. Bjorkman A
    . Complexity of Plasmodium falciparum infections is consistent over time and protects against clinical disease in Tanzanian children. J Infect Dis 1999;179:989-95.
    Abstract/FREE Full Text
  43. 43.↵
    1. Djimde A,
    2. Doumbo OK,
    3. Cortese JF
    . A molecular marker for chloroquine- resistant falciparum malaria. N Engl J Med 2001;344:257-63.
    CrossRefMedlineWeb of Science
  44. 44.↵
    1. Hayes RJ,
    2. Marsh K,
    3. Snow RW
    . Case-control studies of severe malaria. J Trop Med Hyg 1992;95:157-66.
    MedlineWeb of Science
  45. 45.↵
    1. Djimde A,
    2. Doumbo OK,
    3. Steketee RW,
    4. Plowe CV
    . Application of amolecular marker for surveillance of chloroquine-resistant falciparum malaria. Lancet 2001;358:890-1.
    CrossRefMedlineWeb of Science
  46. 46.↵
    Assessment of therapeutic efficacy of antimalarial drugs for uncomplicated falciparum malaria in areas with intense transmission. Division of Control of Tropical Diseases. Geneva: World Health Organization; 1996.
  47. 47.↵
    1. Plowe CV,
    2. Doumbo OK,
    3. Djimde A
    . Chloroquine treatment of uncomplicated Plasmodium falciparum malaria in Mali: parasitological resistance vs. therapeutic efficacy. Am J Trop Med Hyg 2001;64:242-6.
    Abstract
| Table of Contents

This Article

  1. J Infect Dis. (2002) 185 (3): 380-388. doi: 10.1086/338566
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 205 (5)
  1. Journal of Infectious Diseases
  1. Alert me to new issues

Most