Pfmdr1 Alleles and Response to Ultralow-Dose Mefloquine Treatment in Gabonese Patients

Abstract

The identification of parasite molecular markers involved in resistance to antimalarial compounds is of great interest for monitoring the development and spread of resistance in the field. Polymorphisms in Plasmodium falciparum multidrug resistance gene 1 (pfmdr1) have been associated with chloroquine resistance and mefloquine susceptibility. In the present study, carried out in Lambaréné, Gabon, we investigated the relationship between the presence of mutations at codons 86, 184, 1034, 1042, and 1246 in the pfmdr1 gene and the success of ultralow-dose mefloquine treatment (1.1 mg/kg of body weight). Sixty-nine patients were included in the study, and depending on the level of in vivo resistance to mefloquine, they were classified as sensitive responders (S), patients with low-grade resistance (RI), and nonresponders (NR). We found that the prevalences of the Tyr-86 mutation among isolates from patients in groups S, RI, and NR were 100, 96, and 90%, respectively, and that the prevalence of the Phe-184 mutation among the isolates was 80% in each group. A prevalence of about 10% point mutations at codons 1042 and 1246 was detected only in isolates from patients in groups RI and NR. There was no statistically significant association between the presence of the Tyr-86 mutation and the in vivo response (P = 0.79). Among the parasite isolates from patients with drug-resistant infections, 83% had the wild-type pfmdr1 genotype (S₁₀₃₄-N₁₀₄₂-D₁₂₄₆). No link between the presence of this genotype and parasite resistance was detected (P = 0.42). Among the isolates analyzed, 85 had double mutations (Y₈₆-F₁₈₄ or Y₈₆-Y₁₂₄₆) and 11 had triple mutations (Y₈₆-D₁₀₄₂-Y₁₂₄₆, Y₈₆-F₁₈₄-Y₁₂₄₆, or Y₈₆-F₁₈₄-D₁₀₄₂). These findings are not consistent with those of previous in vitro studies and suggest that further evaluation of pfmdr1 gene polymorphism and in vivo mefloquine sensitivity are needed.

The rapid spread of chloroquine resistance has led to the evaluation and development of second-line treatments for uncomplicated malaria (26, 27). In Southeast Asia, alternative drugs such as mefloquine were introduced for the treatment of uncomplicated falciparum malaria, and as early as 1984 resistance to this drug was reported in Thailand (14, 26, 27). In West Africa, resistance to mefloquine has been described from areas where the drug has not previously been used (5, 15).

The molecular mechanisms involved in mefloquine resistance are not fully understood, but an increase in the size of chromosome 5 after exposure to mefloquine has been described and seems to be an important feature. This increase of the chromosome size is caused by an amplification of Plasmodium falciparum multidrug resistance gene 1 (pfmdr1) (7, 16, 28). However, an in vitro study did not show an association between pfmdr1 gene size and mefloquine resistance (1). Moreover, the analysis of field isolates showed that amplification and overexpression of the pfmdr1 gene were not necessary for increased mefloquine resistance (12). Point mutations present in various codons of the pfmdr1 gene such as codons 86 (asparagine [N] → tyrosine [Y]), 184 (tyrosine [Y] → phenylalanine [F]), 1034 (serine [S] → cysteine [C]), 1042 (asparagine [N] → aspartate [D]), and 1246 (aspartate [D] → tyrosine [Y]) have been associated with chloroquine resistance (9). Their contributions to mefloquine resistance are questionable. It has been shown in heterologous expression systems that introduction of these mutations, specifically at codons 1034 and 1042 of the pfmdr1 gene, abolished or reduced the level of resistance to mefloquine (23, 25). Moreover, transfections of the pfmdr1 gene showed that the wild-type allele at codons 1034, 1042, and 1246 can confer mefloquine resistance to sensitive P. falciparum strains (21). Analysis of P. falciparum isolates from The Gambia and Thailand showed an association between the presence of Y₈₆ (mutated type) and increased sensitivity to mefloquine (8, 19).

The present study was undertaken to investigate pfmdr1 gene polymorphisms in P. falciparum parasite populations from Lambaréné, Gabon. In that area the standard regimen of mefloquine showed high levels of efficacy in schoolchildren (20). A clinical trial with ultralow-dose mefloquine for the treatment of children in Gabon (11) gave us the opportunity to test the hypothesis that failure of treatment with ultralow-dose mefloquine is associated with the presence of Y₈₆ or other mutations in the pfmdr1 gene. Furthermore, we compared the breakthrough parasite genotypes with the genotypes that existed before the start of treatment.

MATERIALS AND METHODS

Study site and blood collection.

The study was carried out from January 1995 to January 1996 at the Medical Research Unit of the Albert Schweitzer Hospital in Lambaréné, Gabon. Lambaréné is located in an area hyperendemic for malaria, where the predominant malarial species is P. falciparum.

The study design was previously described in detail (11). Briefly, patients were included in the study if they met the following criteria: (i) the patient was infected only with P. falciparum; (ii) the patient had clinical symptoms and a recent history of fever; (iii) signs of severe malaria, such as severe anemia, cerebral malaria, or hypoglycemia were absent; (iv) other severe coinfections or infections with other Plasmodium species were absent; (v) the patient had no history of recent treatment with antimalarial drugs, confirmed by testing of urine samples for chloroquine and quinine and for sulfonamides; and (vi) the parents or legal guardians provided informed consent. The children were aged 3 to 15 years; and the parasite density, determined with a Giemsa-stained thick blood smear, was in the range of 1,000 and 100,000 parasites/μl. Patients received, on average, 1.1 mg of mefloquine/kg of body weight in a single dose, which corresponds to about 1/12 of the normal dose. Blood smears and clinical examination were done until clinical and parasitological cure and on days 7, 14, 21, and 28. In case of treatment failure, the patients were cured by use of a regular dose of the combination treatment (12.5 mg of mefloquine/kg, 25 mg of sulfadoxine/kg, 1.25 mg of pyrimethamine/kg). The study was approved by the Ethics Committee of the International Foundation of the Albert Schweitzer Hospital.

The patients were classified into three groups according to their clinical responses. (i) A sensitive response (group S) was defined as clearance of parasites during the first week and no recrudescence throughout the follow-up of 28 days. (ii) Low-grade resistance (group RI) was defined as clearance of parasites during the first week with a reappearance of parasites during the follow-up. (iii) Nonresponse (group NR) was defined as a failure to clear the parasites during the first week or as an increase in parasite density 2 days after treatment or later compared to the parasite density at the baseline.

P. falciparum isolates from all patients were analyzed, and a pair of isolates refers to isolates that were collected from the same patient before the start of treatment and at recrudescence.

P. falciparum DNA extraction.

Parasite DNA was extracted with a QIAamp blood DNA kit (Qiagen, Hilden, Germany). Two hundred microliters of parasitized red blood cells was required to perform this extraction, according to the manufacturer’s instructions. The extracted DNA was stored at −20°C until use.

Pfmdr1 gene polymorphism analysis.

Point mutations were detected by different nested PCRs and successive enzymatic digestions. PCR amplifications were performed with the oligonucleotides described by Duraisingh et al. (8) in a final volume of 50 μl in a Perkin-Elmer 480 thermocycler. Each PCR mixture contained 200 μM deoxynucleoside triphosphates, 1.5 U of Taq polymerase (Qiagen), and 1.5 mM MgCl₂ (Qiagen). A negative control without parasite DNA and a positive control containing DNA prepared from laboratory-adapted P. falciparum strain Binh I (3) were used in the different PCRs.

For the identification of the different point mutations, restriction enzymes were used. The ApoI and DraI enzymes cut the wild-type sequence and the mutant sequence specific for asparagine at codon 86 (N₈₆) and phenylalanine at codon 184 (F₁₈₄), respectively. DdeI and AseI were used to detect the wild-type allele of serine at codon 1034 (S₁₀₃₄) and the wild-type allele of asparagine at codon 1042 (N₁₀₄₂), respectively. EcoRV was used to analyze codon 1246 and allowed the detection of the tyrosine (Y₁₂₄₆) mutant allele. As the parasite genome is haploid during the asexual blood stage, a single isolate in which both wild-type and mutant alleles were detected was considered to arise from an infection with multiple strains.

P. falciparum strain Binh I was analyzed for the different point mutations and was used as a control in the digestion reactions. Tyrosine, phenylalanine, serine, asparagine, and aspartate were found at positions 86, 184, 1034, 1042, and 1246, respectively. To verify the results of the enzymatic digestion, samples of nondigested PCR products were sequenced with an ABI prism sequencer (Applied Biosystems, Inc., Foster City, Calif.). PCR and digested products were analyzed on an agarose gel (1 to 3.5%) stained with ethidium bromide and were visualized under UV light.

Statistical analysis.

In each group of patients (groups S, RI, and NR), the proportion of wild-type and mutant amino acids per codon was determined as follows: number of wild-type or mutant codons/total number of alleles (wild type plus mutant) detected. Fisher’s exact test was used to compare the genotypes of the parasites from the groups with the different outcomes. The McNemar test was used to compare the genotypes of the parasites obtained before treatment and at recrudescence. A P value of <0.05 was considered significant.

RESULTS

Of the 76 patients included in the study by Lell et al. (11), only blood samples from 69 patients were available for the present study and a total of 130 isolates were obtained. Different fragments of DNA at all five point mutations were successfully amplified from 111 isolates obtained from 58 patients: 5 isolates from group S, 24 pairs from group RI, and 29 pairs from group NR.

Prevalence of point mutations.

We evaluated the prevalence of the point mutations in P. falciparum isolates selected before treatment from cured patients (group S) and from patients whose infections were persistent (group NR) or recurrent (group RI).

Mutations at positions 86 and 184 were more prevalent than mutations at positions 1034, 1042, and 1246. Indeed, 93 and 82% of mutant types were found at positions 86 and 184, respectively (Table 1). The wild-type allele was predominantly found at codons 1042 and 1246, with 90% being N₁₀₄₂ and 93% being D₁₂₄₆. All the isolates had the wild-type allele at position 1034 (S₁₀₃₄).

TABLE 1.

Prevalences of point mutations in pfmdr1 gene of isolates in the groups of Gabonese patients (groups S, RI, and NR) whose isolates were resistant or sensitive to ultralow doses of mefloquine

Group	Prevalence (no. [%]) of isolates with point mutationsa
	Total	Wt					Mt
		N86	Y184	S1034	N1042	D1246	Y86	F184	C1034	D1042	Y1246
S	5 (100)	0 (0)	1 (20)	5 (100)	5 (100)	5 (100)	5 (100)	4 (80)	0 (0)	0 (0)	0 (0)
RI	24 (100)	1 (4)	5 (20)	24 (100)	20 (83)	21 (88)	23 (96)	19 (80)	0 (0)	4 (17)	3 (12)
NR	29 (100)	3 (10)	5 (17)	29 (100)	27 (93)	28 (97)	26 (90)	24 (83)	0 (0)	2 (7)	1 (3)
Total	58 (100)	4 (7)	10 (18)	58 (100)	52 (90)	54 (93)	54 (93)	48 (82)	0 (0)	6 (10)	4 (7)

^aThe prevalence of each point mutation was evaluated for isolates collected before treatment. Wt and Mt, wild-type and mutant, respectively.

To determine whether the presence of pfmdr1 mutations in isolates before treatment was associated with subsequent treatment failure, we compared the prevalences of these mutations in P. falciparum isolates from patients in groups S, RI, and NR. We found that all parasites collected from the five patients (100%) in group S had the Y₈₆ allele, whereas 96 and 90% of the parasites from patients in groups RI and NR, respectively, had this allele, although no statistically significant association between the presence of this allele and in vivo sensitivity to mefloquine was found (P = 0.79). The same held true for all the other point mutations examined.

To evaluate whether mefloquine treatment selected for certain parasite genotypes, we compared pairs of isolates: one isolate in the pair was obtained before treatment and the other was obtained at recrudescence. We compared the prevalence of mutations at each codon in pairs of isolates from both group RI (24 pairs) and group NR (29 pairs), and we found no statistically significant difference for any codon (P = 0.5). For 14 (26%) pairs of isolates there was a change in the genotype: 9 in group RI and 5 in group NR. Isolates from five subjects had changes at codon 86, isolates from six subjects had changes at codon 184, isolates from six subjects had changes at codon 1042, and isolates from two subjects had changes at codon 1246. Isolates from none of the subjects had changes at codon 1034, and isolates from four subjects had changes at more than one codon. We found no statistically significant selection of wild-type or mutant amino acids during treatment for any codon (P ≥ 0.47 for all codons).

Pfmdr1 sequence polymorphism and resistance to mefloquine.

We analyzed the association between pfmdr1 sequence polymorphism and the response to treatment in the 58 patients. As shown in Table 2, two different sequences were found in isolates from the group of patients whose infections were cured with mefloquine. Seven distinct sequences were observed among isolates resistant to mefloquine and collected before treatment from patients in both group RI and group NR. Among these sequences, a predominant one, Y₈₆-F₁₈₄-S₁₀₃₄-N₁₀₄₂-D₁₂₄₆, was observed in both groups. Before treatment, we found that 39 of 58 isolates (67%) had this genotype. Four of five patients (80%) in group S were infected with parasites with this genotype. Among the isolates from patients in group RI, 62 and 67% of isolates had this genotype before treatment and at recrudescence, respectively, whereas among the isolates from patients in group NR, 69% of the isolates harbored this genotype both before treatment and at recrudescence. As it has been reported that pfmdr1 genotype S₁₀₃₄-N₁₀₄₂-D₁₂₄₆ is associated with mefloquine resistance (21), we analyzed this sequence in more detail. Among the 49 patients harboring parasites with this sequence on admission, 44 (90%) were not cured, whereas 9 (100%) patients harboring parasites with different genotypes on admission were not cured. No statistically significant association between the presence of this sequence and mefloquine treatment failure was observed (P = 0.42).

TABLE 2.

Pfmdr1 gene polymorphisms in isolates from Gabonese patients resistant or sensitive to ultralow doses of mefloquine

Group	Amino acid at the following position before treatmenta:					No. of isolates for before treatmentb	Amino acid at the following position after treatmenta:					No. of isolates for after treatmentb
	86	184	1034	1042	1246		86	184	1034	1042	1246
S	Y	Y	S	N	D	1
	Y	F	S	N	D	4
RI	Y	Y/Fc	S	N	Y	1	Y	Y/Fc	S	N	Y	1
	Y	F	S	N	D	15	Y	F	S	N	D	16
	Y	Y	S	N	Y	1	Y	Y	S	N	Y	1
	Y	F	S	D	D	3	Y	F	S	D	D	2
	Y	Y	S	N	D	2	Y	Y	S	N	D	2
	Y	Y	S	D	Y/Dc	1	N	Y	S	D	D	1
	N	F	S	N	D	1	Y/Nc	F	S	N	D	1
NR	Y	F	S	D	D	2	Y	F	S	D	D	1
	Y	Y	S	N	Y	1	Y	Y	S	N	Y	1
	Y	F	S	N	D	20	Y	F	S	N	D	20
	Y	Y	S	N	D	2	Y	Y	S	N	D	3
	N	Y	S	N	D	1	N	Y	S	N	D	1
	Y/Nc	F	S	N	D	2	Y/Nc	F	S	N	D	1
	Y	Y/Fc	S	N	D	1	Y	Y	S	N	D/Yc	1
							N	F	S	N	D	1

^aThe predominant pfmdr1 genotype Y₈₆-F₁₈₄-S₁₀₃₄-N₁₀₄₂-D₁₂₄₆ is presented in boldface.

^bNumbers in italic represents the number of isolates with the S₁₀₃₄-N₁₀₄₂-D₁₂₄₆ genotype.

^cPresence of multiple P. falciparum infections.

No isolate analyzed had amino acids with mutations at all five codons. Mutants with double mutations at codons 86 and 184 were present in each group of patients, whereas mutants with double mutations at codons 86 and 1246 were detected only in isolates from the patients in groups RI and NR. Among isolates from patients in groups RI and NR we also found isolates with triple mutations at codons 86, 184, and 1042; codons 86, 1042, and 1246; and codons 86, 184, and 1246.

Analysis of mixed P. falciparum infections.

Mixed P. falciparum infections were identified by the detection of both wild-type and mutant amino acids in pfmdr1 in a single isolate. Among the 111 isolates tested, 9 were found to be part of multiple infections (8%). Isolates from patients with multiple P. falciparum infections with mutations at codon 86 (n = 4) were distinct from those with mutations at codon 184 (n = 3) and from those with mutations at codon 1246 (n = 2) (Table 2). No multiple P. falciparum infections were detected among the patients in group S. Three individuals each in groups RI and NR harbored mixed infections.

DISCUSSION

The purpose of the present study was to characterize the pfmdr1 gene polymorphism in isolates from patients with uncomplicated malaria in Gabon and to assess the possible selection of pfmdr1 genotypes following ultralow-dose treatment with mefloquine. Specifically, we have investigated the relationship between the presence of five different mutations in the pfmdr1 genes of P. falciparum isolates and the in vivo response to mefloquine. It is the first in vivo study to investigate this marker for mefloquine sensitivity.

The in vitro response to mefloquine and its efficacy have frequently been assessed in Lambaréné, Gabon (4, 17, 24, 29). From 1994 to 1996 the 50% effective concentration of mefloquine ranged between 510 and 360 nmol/liter below the threshold for resistance, which was 3,200 nmol/liter. The presence of mutations on the pfmdr1 gene was investigated in isolates from patients treated with chloroquine in the same area in Gabon (10), and the results showed a high prevalence (80%) of Y₈₆, whereas no mutation at position 1246 (Y₁₂₄₆) was observed. This high prevalence of mutations at codon 86 was also observed in Cameroonian isolates resistant to chloroquine (2). In the present study, we found that more than 90% of isolates displayed Y₈₆, and among these, only 9% were cleared by low-dose mefloquine. No association was found between the mutation at codon 86 of the pfmdr1 gene and in vivo sensitivity to mefloquine, which is in contrast to the results of in vitro studies (8, 19).

Previous studies have mainly focused on the association of the pfmdr1 gene copy number and in vitro sensitivity to mefloquine (6, 7, 12, 13, 16, 22, 28). Few in vitro studies have examined the relationship between pfmdr1 gene mutations and mefloquine sensitivity, and the studies that have been conducted have obtained controversial results. Two investigations carried out in Thailand and The Gambia showed an association between a mutant allele (Y₈₆) at codon 86 and increased mefloquine sensitivity (8, 19), whereas Zalis et al. (30) described a link between the wild-type allele (N₈₆) and increased mefloquine sensitivity. In the latter study, the pfmdr1 gene was analyzed in 26 isolates from South America which were sensitive to mefloquine in vitro, and all had the N₈₆ wild-type allele.

The relationship between in vivo mefloquine sensitivity and the presence of Y₈₆ on the pfmdr1 gene has not previously been described, and our results show a trend toward an increasing prevalence of Y₈₆ in isolates from the group with a low level of sensitivity to mefloquine (group NR) to the group with a high level of sensitivity (group S). However, these findings, taking into account the small sample size in group S, do not support the hypothesis that the Y₈₆ mutation is an adequate molecular marker for mefloquine sensitivity. However, larger studies in areas with high levels of mefloquine resistance are needed to determine the clinical relevance of the Y₈₆ allele for mefloquine treatment.

The prevalences of mutations at codons 184, 1034, and 1246 in the present study are similar to those reported by Basco et al. (1) in isolates from West Africa. The low prevalence of mutations observed at codon 1246 in African isolates is in contrast to that observed in isolates from South America (18).

In vitro experiments have shown that parasites with wild-type pfmdr1 sequences at codons 1034, 1042, and 1246 together have lower levels of mefloquine sensitivity (21). However, our study could not confirm these findings. Neither the S₁₀₃₄-N₁₀₄₂-D-₁₂₄₆ sequence nor any other sequence was associated with the clinical outcome after ultralow-dose mefloquine treatment.

In conclusion, the results of our study with low-dose mefloquine treatment suggest that the pfmdr1 sequence does not accurately predict in vivo mefloquine sensitivity, contradicting previous in vitro findings. However, the results of in vivo studies do not necessarily correspond to the in vitro findings, since individuals can have different susceptibilities and certain degrees of immunity to the disease. If a practical application is anticipated, in vivo studies are necessary to verify markers of resistance in field situations. The ultralow dose of mefloquine given to patients in the present study did not allow any speculation on the use of pfmdr1 gene point mutations as markers of mefloquine resistance in the field. Pfmdr1 gene polymorphism analysis including the pfmdr1 gene copy number should also be investigated in isolates from patients treated with a normal regimen of the drug to confirm or not confirm our findings.