In Vitro–Reduced Susceptibility to Artemether in P. falciparum and Its Association With Polymorphisms on Transporter Genes

Abstract

Plasmodium falciparum with reduced sensitivity to artemisinin derivatives has been observed in endemic areas, but the molecular mechanisms for this reduced sensitivity remain unclear. We evaluated the association between in vitro susceptibility of P. falciparum isolates obtained from southwest Nigeria and polymorphisms in selected putative transporter genes (PFE0775C, PF13_0271, pfmrp1, pfcrt, and pfmdr1). Modified schizont inhibition assay was used to determine the in vitro parasite susceptibility to artemether (ATH). Polymorphisms in selected genes were detected by polymerase chain reaction followed by direct DNA sequencing. The half-maximal inhibitory concentration (IC₅₀) geometric mean (GM) for all P. falciparum isolates was 1.78 nM (range, 0.03–10.43 nM). Polymorphisms at codons 241, 86, and 76 of PFE0775C, pfmdr1, and pfcrt genes, respectively, were associated with reduced susceptibility to ATH. A new S263P single-nucleotide polymorphism on the PFE0775C gene was also detected in 27% of the isolates. Patient isolates harboring V241L or S263P polymorphisms on the PFE0775C gene showed increased IC₅₀ (GM: 3.08 nM and 1.79 nM, respectively). Plasmodium falciparum isolates harboring mutant Y86 pfmdr1 and P263 PFE0775C alleles showed a 2.5–5.5-fold increase in ATH IC_50. This study shows that polymorphisms on the PFE0775C and pfmdr1 genes are associated with reduced sensitivity to ATH in fresh isolates of P. falciparum from Nigeria.

The emergence of resistant Plasmodium falciparum to all available antimalarial drugs is a major challenge to malaria control in endemic areas [1]. Antimalarial drugs such as chloroquine (CQ) and sulfadoxine-pyrimethamine have become ineffective in the treatment of P. falciparum malaria [2]. Other drugs such as mefloquine (MQ), amodiaquine (AQ), and even quinine (QN) have shown reduced efficacy as well [3].

To overcome the widespread and increasing level of parasite resistance to antimalarial drugs, the World Health Organization (WHO) recommended the use of artemisinin-based combination therapies (ACTs) for the treatment of uncomplicated P. falciparum malaria [4, 5]. On the basis of this recommendation, most malaria-endemic countries worldwide changed their antimalarial treatment protocols to the use of artemether (ATH)-lumefantrine (LUM) and artesunate (AS)-amodiaquine (AQ) as first-line treatment of uncomplicated malaria [6].

Artemisinin derivatives (ARTs), are potent antimalarial drugs with activity against both asexual and sexual stages of P. falciparum [7]. They have the ability to clear parasites and resolve symptoms rapidly compared to other antimalarial drugs [5]. This results in more rapid clinical response and reduction in parasite transmission. However, the short half-life of ARTs enables some parasites to escape drug action during treatment, thus the reason for its combination with other antimalarial drugs with longer half-lives [8].

Despite the advantages of ARTs, reduced efficacy of the AS–MQ combination therapy [9, 10] and in vitro resistance to AS [11] have been reported on the Thai-Cambodia border and in western Cambodia, respectively. Clinical failures to ART monotherapies observed in the past have been ascribed to pharmacokinetic factors, in particular, rapid elimination half-life and inadequate dosage [12]. Laboratory studies have shown that P. falciparum can develop stable resistance to ARTs with high half-maximal inhibitory concentration (IC₅₀)_, suggesting the possible emergence of in vivo resistant parasites after prolonged exposure and extensive use of this drug [13, 14].

The continued availability and use of ARTs as monotherapy for treatment of uncomplicated malaria may subject P. falciparum to drug selection pressure by exposure to subtherapeutic concentrations, especially in areas of intense malaria transmissions of Africa. A recent study showed that ARTs monotherapy were prescribed to 18.2% and 15.8% of malaria patients in southwest and southeast Nigeria, respectively [15].

The mechanism of action and resistance of ARTs is not clearly understood. Some studies suggest that the mechanism of action is related to the scission of the peroxide bridge by reduced haem iron, which is produced inside the highly acidic digestive vacuole as it digests hemoglobin [16]. An alternative mechanism of action of ARTs is the inhibition of a sarcoplasmic/endoplasmic reticulum Ca2⁺-ATPase [17]. This enzyme is important for the generation of calcium-mediated signaling and the correct folding and posttranslational processing of proteins [18].

In P. falciparum isolates from French Guiana, polymorphisms in the PfATPase6 gene were associated with significant increase in IC₅₀ to ATH [19]. However, these polymorphisms were not found in isolates from other disease-endemic countries [20–22]. Polymorphisms on pfmdr1 and P. falciparum multiple resistant associated protein 1 (pfmrp1) genes have also been reported to be associated with reduced susceptibility to ARTs [23]. An elevated pfmdr1 gene copy number has been reported to reduce parasite susceptibility to ARTs [24] and increase the risk of treatment failure with combination of AS and MQ on the Cambodian-Thai border [25]. Unfortunately, not all of these findings are being validated in many malaria-endemic areas, including western Cambodia where artemisinin resistance has emerged [11]. The fact that reduced sensitivity of ARTs is not explained by single-nucleotide polymorphisms on the pfATPase6, pfcrt, pfmd1, or pfmrp1 genes, nor by amplification of these genes, suggests that resistance to ARTs may be multigenic.

We hypothesized that polymorphisms in transporter proteins in P. falciparum may be involved in reduced susceptibility of field isolates to ARTs. Polymorphisms in orthologues of Plasmodium transporters in microorganisms and human tumor cells are known for their contribution to drug resistance. In addition, drug resistance in other human malaria species and the induction of artemisinin (ART) resistance in mouse malaria models has been associated with polymorphisms in the parasite transporter genes.

In this study, the involvement of polymorphisms in selected transporter genes pfrmrp1 (GenBank accession no. GU797311), PFE0775C (GenBank accession no. X_M001351676), and PF13_0271 (GenBank accession no. AL844509), and pfcrt (GenBank accession no. AF233068) and pfmdr1 (GenBank accession no. AL844504) genes with in vitro–reduced susceptibility of P. falciparum isolates to ATH were investigated in fresh patient isolates of P. falciparum in a major clinical testing center in Nigeria. We found an association between polymorphisms in parasite transporter genes and reduced susceptibility to ATH. The potential implications of these polymorphisms on the development of parasites with reduced susceptibility to ATH in endemic areas where ACTs are being used as first-line treatment of uncomplicated falciparum malaria are discussed.

MATERIALS AND METHODS

Study Site and Sample Collection

This study was conducted at the Malaria Research Laboratory clinic, College of Medicine, University of Ibadan, Nigeria, and the Laboratory of Molecular Infectology and Parasitology, Clementino Fraga Filho University Hospital, Rio de Janeiro, Brazil. Children aged 6 months to 12 years with symptoms of uncomplicated P. falciparum malaria were enrolled after clinical examination and microscopic confirmation of infection in a large clinical efficacy study in Ibadan.

Written informed consent for participation in the study was obtained from each patient or parent/guardian. The joint University of Ibadan (UI)/University College Hospital (UCH) Institutional Review Committee approved the study protocol. Each child was treated with a standard dose of AS-AQ and followed up for a 42-day period in accordance with WHO protocol. Thin and thick blood films were prepared from finger-prick blood samples prior to treatment for parasite estimation. Finger-prick blood samples were also spotted on Whatman 3MM filter paper for molecular analysis. Venous blood (5 mL) was collected in sterile heparinized tubes from each patient prior to treatment for ATH in vitro susceptibility testing.

In Vitro Testing

The in vitro susceptibility of each patient isolate to ATH was determined using a modification of the schizont inhibition assay [26]. In brief, a template containing 3-fold serial dilutions of ATH (209.4 nM) was prepared in a 96-well microtiter plate. Wells in row H served as controls without drug. Test plates were derived from each template by transferring 25 µL of the drug dilutions to each plate. Two hundred microliters (200 µL) of 1 mL parasitized blood diluted in 19 mL of culture medium (RPMI 1640 plus HEPES and sodium bicarbonate) was transferred into each well. Plates containing parasite suspension with ATH were incubated at 37°C for 24–36 hours in a plexiglass chamber containing a gas mixture (5% O₂, 5% CO₂, and 90% N₂). The final concentration of ATH in wells A–G ranged from 209.4 nM to 0.3 nM.

The assay was terminated when at least 60% of parasites in the control wells (row H) were schizonts. Each well in a column of 96-well plates was harvested onto glass slides in thick smears, air dried, and stained with Giemsa. Parasite development to schizont was determined by counting the number of schizonts against 200 white blood cells in each smear using a light microscopy oil immersion objective (×100). Concentration-response data were analyzed by a nonlinear regression analysis. The IC₅₀ for ATH was calculated using GraphPad Prism version 4 software.

Molecular Analysis

DNA Extraction and Nested Polymerase Chain Reaction

Parasite genomic DNA was extracted from filter paper containing blood samples using phenol-chloroform with saponin lysis method [27].

Nine different loci in 5 different genes that encode transporter proteins were chosen for amplification and sequencing (Table 1). These genes have been previously associated with in vitro reduced susceptibility to CQ and QN in P. falciparum [28]. Nested polymerase chain reaction (PCR) was used for detection of polymorphisms on the pfcrt [29], pfmdr1 [30], pfmrp1, PF13_0271, and PFE0775C genes. Primer sequences and reaction conditions for the genes are shown in Table 2. All PCR reactions were performed in a final volume of 25 µL in a final concentration of 1 × PCR buffer, 250 µ of each deoxy-nucleotide triphosphate (dNTP), 50 pM each of forward and reverse primers, 1.5 mM MgCl₂, and 2 U of Taq polymerase. Two microliters of the parasite genomic DNA was used as a template for primary amplification and 1 µL of the primary amplification product was used as a DNA template for secondary amplification.

Table 1.

Selected Loci in 5 Different Genes for Polymerase Chain Reaction and Sequencing

Gene ID	NCBI RefSeq	Chromosome Location	Predicted Product	Loci	AA Change
MAL7P1.27 (pfcrt)	NC_004328.2	7	Putative transporter	72–76	CVMNK > CVMNT or CVIET
PFE1150w (pfmdr1)	NC_004326.1	5	ABC transporter	86	N > Y
PFA0590w (pfmrp)	NC_004325.1	1	ABC transporter	437	S > A
PF13_0271	NC_004331.2	13	ABC transporter	890	delN
PFE0775C	NC_004326.1	5	Amino acid transporter	241	V > L
				263	S > P

Abbreviations: AA, amino acid; NCBI RefSeq, National Center for Biotechnology Information Reference Sequence.

Table 2.

Primer Sequences and Conditions for Amplification Reactions of Selected Plasmodium falciparum Genes

	Primary Amplification		Secondary Amplification
PlasmoDB Link	Primer Names and Sequences	PCR Conditions	Primers Names and Sequences	PCR Conditions
PFA0590w	G2bEx: 5′ TATTTATAATATTATGTTTTC 3′	95°C/3 min; 95°C/30 s	G2bInF: 5′ CAATGATACTATTTGAATTT 3′	95°C/3 min; 95°C/30 s:
	G2bExR: 5′ TTTCTTCTTTCTTATTTAATC 3′	49°C/30 s; 70°C/30 s:	G2bInR: 5′ CTTATTTAATCTATCTTTTA 3′	49°C/30 s; 70°C/30 s:
		35 cycles; 70°C/10 min		35 cycles; 70°C/10 min
PF13_0271	G7ExF: 5′ GTAATGTGAAGAATATCTCA 3′	95°C/3 min; 95°C/30 s:	G7InF: 5′ CAAATCCAAATATTACGAAAA 3′	95°C/3 min; 95°C/30 s:
	G7ExR: 5′ TTGAAGCTTGAATCATTTGTTTATC 3′	50°C/30 s; 70°C/30 s	G7InR: 5′ AGTATCTTGTGGTACGACACTT 3′	50°C/30 s; 70°C/30 s:
		35 cycles; 70°C/10 min		35 cycles; 70°C/10 min
PFE0775C	G47ExF: 5′ GTATAGATATTAAAGATGCC 3′	95°C/3 min; 95°C/30 s:	G47InF: 5′ GATGCCAAAGAAAAAGAACGA 3′	95°C/3 min; 95°C/30 s:
	G47ExR: 5′ CATATTTTCAAATACACTCGCCATA3′	55°C/30 s; 70°C/30 s:	G47InR: 5′ GACCAGAAGAATGAAATACATCCA 3′	55°C/30 s; 70°C/30 s:
		35 cycles; 70°C/10 min		35 cycles; 70°C/10 min
pfcrt	CRT1: 5′ CGGTTAATAATAAATACACGCAG 3′	95°C/3 min; 95°C/30 s:	CRD1: 5′ TGTGCTCATGTGTTTAAACTT 3′	95°C/3 min; 95°C/30 s:
	CRT2: 5′ CGGATGTTACAAAACTATAGTTACC3′	56°C/30 s; 60°C/60 s:	CRD2: 5′ CAAAACTATAGTTACCAATTTTG 3′	48°C/30 s; 65°C/30 s:
		45 cycles; 60°C/5 min		30 cycles; 65°C/5 min
pfmdr1	MDR1: 5′ CGCGCGTTGAACAAAAAGAGTACCGCTG 3′	95°C/5 min; 95°C/30 s:	MDR3: 5′ TTTCCGTTTAAATGTTTACCTGC 3′	95°C/5 min; 95°C/30 s:
	MDR2: 5′ GGGCCCTCGTACCAATTCCTGAACTCAC 3′	45°C/30 s; 65°C/45 s:	MDR4: 5′ CCATCTTGATAAAAAACACTTCTT 3′	45°C/30 s ; 65°C/45 s:
		45 cycles; 72°C/5 min		45 cycles; 72°C/5 min

DNA Sequence Analysis

The PCR amplification products for each gene were purified using Wizard SV Gel and PCR Clean-Up System kit (Promega) and sequenced directly on an ABI PRISM 3100-Avant Genetic Analyzer with BigDye Terminator v3.1 Cycle Sequencing kit according to the manufacturer's instructions.

Data and Statistical Analysis

Mean, confidence interval, and sample range of the IC₅₀ were calculated using XLSTAT version 2007, and Excel add-in (Addinsoft, 2007). Sequence analysis was performed using FinchTV (Geospiza) for sequence chromatogram visualization, and mutations were localized using Mutation Surveyor (Softgenetics). The 3D7 DNA sequence from PlasmoDB database (http://plasmodb.org/plasmo/) served as reference. We tested the equality of population medians of the in vitro tests among polymorphisms on the genes using the nonparametric Kruskal–Wallis 1-way analysis of variances by ranks implemented as function “kruskal.test” in the R statistical package (R Core Development Team, 2009). Associations between IC₅₀ and polymorphisms on the genes were analyzed by fitting a regression tree. The partitioning methods used have been described previously [31] and implemented in function “rpart” [32] for the R package. In our call to rpart, we used a complexity parameter equal to 0.01 (the default), which means that if any split does not increase the overall R2 of the model by at least complexity parameter (where R2 is the usual linear-model definition), then that split is not worth pursuing. The correlation between in vitro tests and polymorphisms on the genes of interest was determined using Kruskal–Wallis analysis. P < .05 was considered statistically significant.

RESULTS

In Vitro Susceptibilities of P. falciparum Isolates

Plasmodium falciparum isolates were obtained for in vitro susceptibility to ATH from 109 children with uncomplicated malaria. In vitro susceptibility testing was successful in 101 (93%) isolates. The GM IC₅₀ for all the isolates was 1.78 nM (95% confidence interval, 1.42–2.15; range, 0.03–10.43 nM) (Figure 1).

Figure 1.

A, Distribution of artemether IC₅₀ of Plasmodium falciparum isolates with geometric half-maximal inhibitory concentration (IC₅₀) mean of 1.78 nM (horizontal line) and box plot (B) of artemether IC₅₀ with median of 1.07 nM (dark horizontal line), interquartile ranges (gray boxes), and upper and lower adjacent values, which represent the extreme values. The small circles and the asterisks represent outliers and extreme outliers, respectively.

Polymorphisms on pfcrt, pfmdr1, pfmrp1, PFE0775C, and PF13_0271 Genes

Pfcrt Gene

Direct DNA sequencing of amplified products for the pfcrt gene was performed to detect polymorphisms on codons 72–76. Sequencing was successful in 92 P. falciparum genomic DNA samples. The K76T mutation was more predominant among the isolates as it was observed in 73% (67) of the isolates.

Moreover, results showed 3 distinct pfcrt haplotypes: CVMNK, CVMNT, and CVIET. The CVMNK haplotype was observed in 27% (25) of the P. falciparum isolates. The CVMNT and CVIET haplotypes associated with P. falciparum CQ resistance [33] were observed in 3% (3) and 70% (64) of the isolates, respectively.

Pfmdr1 Gene

Nested PCR followed by direct DNA sequencing was also used to detect the presence of N86Y polymorphism on the pfmdr1 gene. One hundred and six (106) isolates were successfully sequenced. The mutant Y86 pfmdr1 allele was observed in 42% (44) of the isolates whereas 58% (62) of the isolates harbored the wild-type N86 pfmdr1 allele.

PF13_0271 Gene

Deletion of asparagine (Asn) on codon 834 on transporter gene PF13_0271 has previously been reported and associated with CQ and QN reduced susceptibility [28]. DNA sequencing was successful in 78 isolates and showed Asn deletion on codon 834 of the gene in 47% (37) of the isolates. The remaining 53% showed no Asn deletion on this gene.

PFE0775C Gene

DNA sequencing of the glycine transporter gene PFE0775C in isolates was successfully performed in 101 isolates of P. falciparum for the detection of V241L polymorphism. Twelve percent of the isolates harbored the V241L polymorphism. In addition to the V241L polymorphism, a new S263P mutation was also detected where proline replaced serine on codon 263 of the gene. This polymorphism was observed in 25% (26) of the isolates sequenced. Two samples analyzed harbored both V241L and S263P mutations.

Pfmrp1 Gene

The P. falciparum multidrug resistance–associated protein 1 (pfmrp1) is a member of the ABC transporter family [34, 35]. One hundred DNA samples were successfully sequenced to detect polymorphisms on the pfmrp1 gene. Of the 100 P. falciparum isolates analyzed, 3% (3) harbored the S437A mutation.

Association Between In Vitro ATH Susceptibility and Genetic Polymorphisms

Association analysis was performed between the presence of polymorphisms on the genes analyzed and in vitro ATH susceptibility. There was no association between parasite in vitro susceptibility to ATH and polymorphisms on PF13_0271 gene. Many P. falciparum isolates, regardless of their susceptibility to ATH in vitro, harbored mutant PF13_0271 alleles. The mutant pfcrtT76 allele was significantly associated with reduced susceptibility to ATH in vitro (P = .036), and the mutant pfmdr1Y86 allele also showed association with decreased in vitro ATH susceptibility (P < .001). Both V241L and S263P mutations in the PFE0775C gene also showed significant association with ATH in vitro susceptibility (P = .01 and P < .001, respectively) (Table 3).

Table 3.

Correlation Between Polymorphisms and Artemether Half-Maximal Inhibitory Concentration

IC50 Geometric Mean (95% CI) and No. of Samples
Gene Polymorphism	Wild-Type	Mutant-Type	P Value
Pfcrt K76T	1.25 (1.25–1.88), n = 19	2.00 (1.53–2.48), n = 68	.03
Pfmdr1 N86Y	1.22 (.87–1.37), n = 50	2.55 (1.89–3.21), n = 48	<.001
PFA0590w S437A	1.73 (1.37–2.09), n = 90	1.79 (.44–3.13), n = 3	.28
PF13_0271 Asp Del	1.76 (1.10–2.42), n = 39	1.90 (1.19–2.62), n = 35	.39
PFE0775C V241L	1.61 (1.27–1.94), n = 86	3.86 (1.76–5.95), n = 10	.01
PFE0775C S263P	1.27 (.98–1.55), n = 68	3.25 (2.30–4.20), n = 28	.001

Abbreviations : CI, confidence interval ; IC₅₀, half-maximal inhibitory concentration.

Correlation Between the Combination of Genetic Polymorphisms and ATH In Vitro Susceptibility

Correlation analysis between combination of polymorphisms on the genes analyzed and P. falciparum isolates’ in vitro sensitivity to ATH was performed and is represented in a regression tree (Figure 2). Plasmodium falciparum isolates harboring the S263P and V241L polymorphisms on the PFE0775C gene were observed to play a significant role on reduced susceptibility to ATH in vitro. Fifteen of 18 samples with the S263P mutation showed higher IC₅₀ (mean = 4.07 nM) compared with the wild-type allele (1.79 nM). Two of these 15 samples harbored both V241L and S263P mutations on the PFE0775C gene and showed high IC₅₀ (4.2 nM and 9.76 nM, respectively). The mean IC₅₀ increased from 0.45 nM to 3.08 nM in the presence of the V241L polymorphism on the PFE0775C gene, whereas the S263P mutation had less influence on the parasite's susceptibility to ATH (1.79 nM). However, the combination of PFE0775C S263P and pfmdr1Y86Y alleles produced the highest mean IC₅₀ (4.07 nM) among the isolates. Seventy-five percent of the samples (12 of 16) with ATH IC₅₀ > 3.3 nM harbored both PFE0775C P263 and pfmdr1Y86 alleles. Two samples harbored both V241L and S263P polymorphisms on the PFE0775C gene and N86Y mutation on the pfmdr1 gene. The IC₅₀ values of these 2 samples were 4.2 nM and 9.76 nM, respectively (ie, 2.36- and 5.49-fold higher than the mean IC₅₀ of all samples). Isolates harboring the combination of N86Y mutation and Asn deletion on pfmrp1 genes respectively showed a mild increase in the mean IC₅₀ compared to isolates without these 2 mutations. The pfcrt K76T mutation alone was able to increase the mean IC₅₀ by 2-fold (from 0.454 to 0.939 nM). Eight samples with no mutation on any genes analyzed in this study displayed the lowest IC₅₀ (0.454 nM). No association was found between mutant pfmrp1 and in vitro reduced susceptibility to ATH.

Figure 2.

Regression tree analysis. The tree represents the genes with their respective mutations that were statistically associated with artemether (ATH) susceptibility. 1 represents the gene name, 2 the allele found in the sequence analysis, 3 the mean half-maximal inhibitory concentration (IC₅₀) to ATH among the isolates, and 4 the number of samples with given alleles. The combination of S263P mutation in PFE0775C gene with N86Y mutation in pfmdr1 gene had a higher impact in reducing the mean IC₅₀ of parasites to ATH. Eighteen samples harboring both mutations displayed a mean IC₅₀ of 4.07 nM. Patient isolates harboring the S263P mutation without the pfmdr1N86Y mutation displayed an IC₅₀ of 1.79 nM. The mutation V241L in the PFE0775C gene was able to increase the ATH IC₅₀ to 3.08 nM in parasites harboring this mutation. Eight samples were wild-type for all genes analyzed and these samples displayed the lowest IC₅₀s in the analysis (mean IC₅₀ = 0.454 nM). Asparagine deletion in PF13_0271 in combination with the N86Y mutation showed a mild increase in parasite IC₅₀ to ATH.

DISCUSSION

Plasmodium falciparum resistance to artemisinin was recently reported [9, 11] and has been characterized by reduced susceptibility to both ARTs and ACTs and a prolonged parasite clearance time. These reports have activated interest in identifying molecular markers of resistance to ARTs and in the monitoring of the emergence of parasites resistant to this class of drugs in malaria-endemic areas.

In the present study, we found some Nigerian isolates of P. falciparum with reduced susceptibility to ATH. Reduced susceptibility in these isolates is associated with polymorphisms in selected putative transporter genes.

The presence of parasites with reduced susceptibility to ATH in Nigeria is particularly worrisome as artemether-lumefantrine (Coartem®) or AS-AQ combinations were only introduced for treatment of malaria in 2005 [15]. Plasmodium falciparum with high IC₅₀ to artemisinin and ATH (20.1 nM and 21.4 nM, respectively), was recently isolated in a traveler returning from Nigeria, who took AS prophylactically (two tablets of 50mg each, weekly for 4 weeks) [36]. The reduced susceptibility to ATH observed in this study may not be unconnected to the pressure exerted on the parasites by the use of ARTs monotherapy for treatment of uncomplicated malaria. Such a practice has the potential of subjecting the parasite population to drug selection pressure by exposure to subtherapeutic concentrations, especially in areas of intense malaria transmission like Ibadan, Nigeria. A recent study showed that ARTs were used as monotherapy in 18.2% and 15.8% of prescriptions in southwest and southeast Nigeria, respectively [15]. In addition, the confirmation of the effect of ARTs selection pressure on P. falciparum reduced susceptibility have also recently been demonstrated by the in vitro development of stable highly resistant ARTs parasites [14]. Although there is no consensus among the malaria community on the in vitro cut-off value for ATH resistance, the GM IC₅₀ of ATH in vitro (1.78 nM) in our study is relatively lower compared with isolates obtained from other African countries such as Senegal (mean IC₅₀ = 3.43 nM; range, 0.8–15.2 nM), but similar to those observed in studies from the Philippines, French Guiana, and the Democratic Republic of São Tomé and Príncipe [37–39].

Antimalarial drug (including ARTs) resistance/reduced susceptibility is often associated with altered gene expression or mutations, generally in genes encoding known and established transporter proteins [40–43]. In our study, the pfcrt K76T mutation showed association with decreased susceptibility to ATH (P = .03). In addition, the regression tree analysis showed a 2-fold increase in ATH mean IC₅₀ of parasites harboring this mutation, although the mean IC₅₀ remained low. This implies that although the K76T mutation increases IC₅₀ values in Nigerian P. falciparum isolates, it is not sufficient to significantly reduce parasite susceptibility to ATH.

ART susceptibility has been reported to be influenced by changes in the pfmdr1 gene [44]. The pfmdr1 N86Y mutation was associated with increased ART sensitivity in a study with Gambian isolates [45], in parasite laboratory lines [46], and isolates from Thailand [47]. Unlike these previous studies, we showed an association between the N86Y mutation and increased ATH IC₅₀ (P < .001). This discrepancy could be explained by previous selection of the N86Y mutation as a result of extensive CQ pressure in Nigeria. The association between ART and the N86Y mutation remains controversial, thus, we hypothesized that pfmdr1 could be acting as a biased, geographic marker or a neutral marker in our study, with no substantial influence in the ATH IC₅₀ levels.

In a study with isolates derived from the Thai-Burma (Myanmar) border, Anderson and colleagues [48] found that mutations in the P. falciparum PF13_0271 gene were associated with AS reduced susceptibility. Moreover, a recent study showed that pfmrp1 additionally influences ART susceptibility in vitro. However, in our study, the PF13_0271 Asn deletion was present in almost half of the samples (47.4%) and was not associated with in vitro drug response. The pfmrp1 S437A mutation showed very low frequency among the isolates and therefore no evidence of association between the mutation and ATH IC₅₀. However, PFE0775C gene mutations were associated with decrease in ATH susceptibility. The PFE0775C V241L mutation was first described in 2003 [28] and was associated with CQ and QN response level. In our study, V241L displayed a 2.4-fold shift in mean ATH IC₅₀ and was associated with ATH decreased susceptibility (P = .01). We also found a new S263P mutation in the PFE0775C gene in isolates from Nigeria resulting in a 2.56-fold increase in ATH IC₅₀ and associated with decreased susceptibility (P = .001). Of particular interest are 2 isolates harboring both PFE0775C gene mutations (V241L and S263P) and displaying 2.35-fold and 5.5-fold increases in IC₅₀ values compared to the overall mean ATH IC₅₀. Currently, little is known about the PFE0775C gene. The gene is located on chromosome 5 and contains 3549 bp encoding a 960 amino acid protein. This protein is a member of the neurotransmitter Na⁺ symporter (NSS) family and is conserved among Plasmodium species [49]. It remains to be established if this NSS protein has any interaction with ARTs or is involved somewhat in facilitating any biological or transport process involving ARTs. Although Mu and colleagues [28] were able to associate this gene with reduced susceptibility to CQ and QN, they did not use ARTs in their experiments. Anderson and colleagues [48] did not find an association between reduced susceptibility to AS in vitro and V241L on the PFE0775C gene in isolates from the Thai-Burma border. These previous findings and the results from this study suggest that parasites may have different backgrounds and the mutations could be selected through different treatment policies established in different countries; thus, polymorphisms on the PFE0775C gene could be used as geographical markers.

The regression tree analysis shows that there is no single major marker causing a dramatic impact on IC₅₀ value to ATH like the pfcrt K76T mutation does for CQ susceptibility. Instead, the combination and accumulation of mutations could be playing important roles in decreasing the parasite's susceptibility to ATH, suggesting that resistance to ATH may involve multiple genetic events in Plasmodium falciparum. It is also possible that the polymorphism observed in these 2 genes might have been a result of previous drug selection pressures. We have previously shown that mutations in P. falciparum resulting from CQ selection pressure were involved in AQ resistance in the same study site [50]

Although the resistance mechanism for ARTs remains unknown, parasites with reduced susceptibility to ARTs from different malaria-endemic areas are being reported. To our knowledge this is the first study in the field reporting the association between mutations in the PFE0775C gene and reduced susceptibility to ATH in vitro. More studies, especially in Africa, are needed to validate this finding.