In Vivo Selection of Plasmodium falciparum Pfcrt and Pfmdr1 Variants by Artemether-Lumefantrine and Dihydroartemisinin-Piperaquine in Burkina Faso

Abstract

Plasmodium falciparum Pfcrt-76 and Pfmdr1-86 gene polymorphisms were determined during a clinical trial in Burkina Faso comparing the efficacies of dihydroartemisinin-piperaquine (DHA-PPQ) and artemether-lumefantrine (AL). Significant selection of Pfcrt-K76 was observed after exposure to AL and DHA-PPQ, as well as selection of Pfmdr1-N86 after AL but not DHA-PPQ treatment, suggesting reverse selection on the Pfcrt gene by PPQ. These results support the rational use of DHA-PPQ in settings where chloroquine (CQ) resistance is high.

TEXT

Emergence of Plasmodium falciparum resistance to artemisinin-based combination therapy (ACT) in Asia (1, 2) represents a major threat to the recent gains in malaria control efforts. Partner drugs in ACTs increase the elimination of late-clearing parasites; however, if resistance develops against these partner drugs, treatment failures will probably increase (3).

Artemether-lumefantrine (AL) is currently the most widely used ACT in Africa (4), and dihydroartemisinin-piperaquine (DHA-PPQ) was recently recommended by the WHO for treatment of uncomplicated falciparum malaria (5). Piperaquine (PPQ) monotherapy was extensively used in China until 1978, when it was abandoned due to widespread resistance (6). PPQ has a long half-life of 3 to 4 weeks; therefore, its use may expose reinfecting and late-clearing parasites to a suboptimal blood concentration of PPQ, potentially leading to selection of resistant strains (6).

A P. falciparum chloroquine (CQ) resistance transporter gene mutation at codon 76 (Pfcrt-K76T) has been associated with CQ and amodiaquine (AQ) resistance (7, 8). Due to structural similarities between PPQ and CQ, there have been attempts to identify common markers of resistance (9, 10). However, the limited number of studies available did not observe selection of Pfcrt single nucleotide polymorphisms (SNPs) (11, 12).

Mutations in P. falciparum multidrug resistance gene 1 at codon 86 (Pfmdr1-N86Y) are associated with altered response to structurally unrelated antimalarials, including 4-aminoquinolines and aryl-aminoquinolines (13, 14). Significant selection of Pfmdr1-N86 has been consistently reported with AL in several settings (11, 12, 15). However, there is so far no in vivo evidence of Pfmdr1 selection after DHA-PPQ treatment, with the exception of the reported borderline selection of Pfmdr1-D1246 (12).

In Burkina Faso, artesunate-amodiaquine (AS-AQ) and AL were introduced in 2005 as first-line treatments of uncomplicated malaria (16) following high CQ resistance (CQR) (17). As part of a multicentric study on the safety and efficacy of DHA-PPQ, a clinical trial was conducted in Burkina Faso to test the noninferiority of DHA-PPQ compared to AL (18). We present here Pfcrt and Pfmdr1 SNP analysis of P. falciparum recurrent infections in an attempt to assess whether DHA-PPQ selected for known polymorphisms in this area of known high CQR

A total of 301 children of ages 6 to 59 months were randomly allocated to either DHA-PPQ or AL treatment at a ratio of 2:1 and followed up for 42 days as published elsewhere (18). Blood spot filter paper specimens were collected on the day of enrollment and during follow-up. Genotyping to distinguish recrudescent from new infections was done as previously described (19). The detection of the Pfcrt-76 and Pfmdr1-86 SNPs was carried out by PCR followed by restriction fragment length polymorphism (RFLP) as previously described (7). A total of 272 (91 AL versus 181 DHA-PPQ samples) day 0 samples as well as 24 and 37 samples for recurrent P. falciparum infections in the AL and DHA-PPQ groups, respectively, were analyzed. The χ² test with Yates correction or the Fisher exact test was used for categorical variables as appropriate. The strength of association between markers and treatment outcomes was evaluated by odds ratios (OR). Mixed alleles were treated as mutant, while unsuccessful PCR results were excluded from the analysis. Statistical significance was set at ≤0.05.

Totals of 267 (98.2%) and 272 (100%) day 0 blood samples were successfully genotyped for the Pfcrt-K76T and Pfmdr1-N86Y polymorphisms, respectively (Fig. 1). At baseline, the prevalences of the Pfcrt-76T mutants were similar between the two study arms: about 50% (48.3% for AL and 48.9% for DHA-PPQ) when considering infections with only the mutant alleles and almost 70% when considering infections with mixed alleles (Table 1). The prevalences of the Pfmdr1-N86 allele at baseline were higher: 59.3% for AL and 64.1% for DHA-PPQ.

FIG 1.

Trial profile indicating the number of samples analyzed at day 0 and among recurrences by treatment arm. The treatment outcomes are all PCR corrected.

TABLE 1.

Selection of Pfcrt and Pfmdr1 SNPs among P. falciparum recurrent infections by AL and DHA-PPQ

Treatment and allelea	SNPb	No. of infections/total:		χ2 value	P valuec	Comment
		Day 0	Recurrent
AL
Pfcrt	K76	27/89 (30.3)	19/24 (79.2)
	76T	43/89 (48.3)	1/24 (4.2)	16.7	<0.001	Selection for K76
	Mixed (K76T)	19/89 (21.4)	4/24 (16.7)
Pfmdr1	N86	54/91 (59.3)	23/24 (95.8)
	86Y	17/91 (18.7)	1/24 (4.2)	NAd	<0.001	Selection for N86
	Mixed (N86Y)	20/91 (22.0)	0/24 (0)
DHA-PPQ
Pfcrt	K76	52/178 (29.2)	25/37 (67.6)
	76T	87/178 (48.9)	9/37 (24.3)	18.0	<0.001	Selection for K76
	Mixed (K76T)	39/178 (21.9)	3/37 (8.1)
Pfmdr1	N86	116/181 (64.1)	24/37 (64.9)
	86Y	32/181 (17.7)	9/37 (24.3)	0.0	0.929	No evidence
	Mixed (N86Y)	33/181 (18.2)	4/37 (10.8)

^aPfcrt is the P. falciparum chloroquine resistance transporter gene, and Pfmdr1 is the P. falciparum multidrug resistance gene.

^bFor Pfcrt, K76 is wild type and 76T is mutant, and for Pfmdr1, N86 is wild type and 86Y is mutant.

^cSignificant values are in boldface.

^dNA, not applicable.

The prevalence of the Pfcrt-K76 was significantly higher in recurrent infections than at day 0, before treatment, both in the AL arm (79.2% versus 30.3%; P < 0.001) and in the DHA-PPQ arm (67.6% versus 29.2%; P < 0.001). Similarly, after treatment with AL, the prevalence of Pfmdr1-N86 in recurrent infections was significantly higher than that at day 0 (95.8% versus 59.3%; P < 0.001), while there was no difference in the DHA-PPQ arm (Table 1). No association between Pfcrt-K76T or Pfmdr1-N86Y SNPs and treatment outcome was observed (Table 2).

TABLE 2.

Association between SNPs in Pfcrt-76 and Pfmdr1-86 in P. falciparum isolates and treatment outcomes with AL and DHA-PPQ

Treatment and allelea	SNP	No. of samples with treatment outcome/total (%)b		OR (95% CI)c	P value
		Failed	Cured
AL
Pfcrt	K76	2/7 (28.6)	25/82 (30.5)	0.91 (0.20–6.04)	0.92
	76T	5/7 (71.4)	57/82 (69.5)
Pfmdr1	N86	7/7 (100.0)	49/84 (58.3)	NAd	NA
	86Y	0/7 (0.0)	35/84 (41.7)
DHA-PPQ
Pfcrt	K76	3/13 (23.1)	49/165 (29.7)	0.71 (0.19–2.69)	0.62
	76T	10/13 (76.9)	116/165 (70.3)
Pfmdr1	N86	8/13 (61.5)	109/167 (65.3)	0.85 (0.27–2.72)	0.79
	86Y	5/13 (38.5)	58/167 (34.7)

^aPfcrt is the P. falciparum chloroquine resistance transporter gene, and Pfmdr1 is the P. falciparum multidrug resistance gene.

^bThe denominators represent number of samples for which the SNP was analyzable. Mixed alleles of the Plasmodium falciparum Pfcrt gene or Pfmdr1 gene were treated as mutants. PCR-adjusted treatment failure was defined according to the World Health Organization (27).

^cCI, confidence interval.

^dNA, not applicable.

Our results showed a high baseline prevalence of the Pfcrt-76T alleles, which corroborates previous reports from Burkina Faso (8). Interestingly, significant selection of the Pfcrt-K76 allele after DHA-PPQ treatment was unexpectedly observed as this is contrary to the hypothesis that PPQ and CQ share similar mechanisms of resistance given their structural similarities (20). In comparison, DHA-PPQ did not select for Pfcrt (K76) polymorphisms in previous studies carried out in Burkina Faso (11, 12). The reasons for the observed differences are unknown. However, an earlier in vitro study from Kenya (10) found that PPQ 50% inhibitory concentration (IC₅₀) values were not associated with Pfcrt or Pfmdr1 mutations. Conversely, Pfcrt CQR haplotypes and a novel Pfcrt mutation (C101F) were associated with resistance to PPQ in other studies (14, 20). Overall, the available evidence suggests high in vitro susceptibility of the CQR parasites to PPQ (9, 10, 21). This could be explained by the large bis-quinolone structure of PPQ postulated to inhibit the transporter-mediated drug efflux, thus maintaining high potency against CQR strains (22). Indeed, data from clinical trials confirm the outstanding efficacy of DHA-PPQ despite high CQR levels (18, 23). Thus, as DHA-PPQ becomes increasingly available for treatment and chemoprevention in settings where malaria is endemic, Pfcrt polymorphism and its clinical implications should be further monitored.

Unlike recently reported results from Uganda (24), there was no significant selection of the Pfmdr1-86Y allele observed in our study after DHA-PPQ treatment, consistent with previous reports from Burkina Faso (11, 12). These observed differences may be explained by the different genetic backgrounds of the parasites in West and East Africa. Our results further confirmed that AL treatment selects for Pfcrt-K76 and Pfmdr1-N86 as previously observed (15, 25). At present, AL remains efficacious, although the selected wild-type alleles have been repeatedly associated with diminished sensitivity to lumefantrine (12, 15, 26).

In conclusion, we observed significant selection of Pfcrt-K76 following DHA-PPQ and AL treatment, suggesting that, despite structural similarities between PPQ and CQ, the drugs exert a different mechanism of selection on the Pfcrt gene. These results support the rational use of DHA-PPQ in settings where CQ resistance is high.