Introduction

Malaria is one of the most important public health problems worldwide despite considerable efforts throughout the last century to eradicate or control the disease. Multidrug resistant Plasmodium falciparum, including resistance to structurally related antimalarial aminoquinolines such as chloroquine (CQ), quinine (QN), and mefloquine (MQ), is especially prevalent at the border areas of Thailand (Wongsrichanalai et al. 2002). The development and spread of drug resistance involves primarily the parasite, drug, and human host factors (Hastings et al. 2007). Due to the rapid spread of multidrug resistant falciparum malaria, antimalarial treatment policies in Thailand have been revised periodically. CQ was introduced for clinical use during the period 1965-1974 and followed thereafter by Fansidar™ (sulfadoxine/pyrimethamine), QN, and MQ, during the periods 1974-1980, 1980-1986, and 1986-1995, respectively (Wongsrichanalai et al. 2001). At present, the combination therapy of MQ and an artemisinin derivative artesunate is being used as a first-line treatment for uncomplicated falciparum malaria throughout the country (Malaria-Division 1998).

The association between gene mutation and/or amplification and resistance of P. falciparum to antimalarials has been controversial and debatable during the past decades, depending on geographical areas reported. CQ resistance has been linked to mutations in the pfmdr1 and pfcrt genes which encode the digestive food vacuole transmembrane proteins Pgh1 and PfCRT, respectively (Reed et al. 2000; Fidock et al. 2000). Transfection studies with pfmdr1 indicate that mutation in Pgh1 may modulate the CQ resistance phenotype in vitro; however, in vivo studies showed inconsistent observations (van Es et al. 1994; Zakeri et al. 2008). More recently, a number of point mutations in the pfcrt gene have been linked with CQ resistance (Durrand et al. 2004). A mutation at amino acid position 76 (K76T) was present in all resistant isolates and was thus proposed as a molecular marker of CQ-resistant falciparum malaria (Djimde et al. 2001).

The role of pfmdr1 mutation in conferring reduced susceptibility of the parasite to QN has been proposed. In a South America study (Sidhu et al. 2005), CQ-resistant strains carrying pfmdr1 mutations (N86, C1034, N1042, Y1246) showed a low susceptibility to QN (Mu et al. 2003). In a Gambian study (Zalis et al. 1998), N86Y was weakly associated with decreased QN sensitivity. A molecular study has also implicated the link between pfmdr1 N86Y variant and increased sensitivity of parasites to artemisinin (Pickard et al. 2003). Investigations on the association between pfmdr1 gene copy number as well as polymorphisms of the pfmdr1 gene as possible molecular markers of MQ resistance showed contradictory results (Price et al. 2004). Polymorphisms of pfmdr1 were found to be associated with MQ resistance in some but not all studies. A higher number of gene copy in MQ-resistant isolates was observed in the two studies in Thailand (Price et al. 2004; Nelson et al. 2005; Alker et al. 2007), South America (Sidhu et al. 2005), and Africa (Basco et al. 1995).

In the present study, we investigated the patterns of mutation and amplification of pfcrt and pfmdr1 genes in P. falciparum strains collected from Thai-Cambodian border during two periods, i.e., 1988-1993 and 2003.

Material and methods

Culture system for parasite maintenance

P. falcifarum field isolates were collected from malaria endemic areas of Thailand along Thai-Cambodian border (Trad and Chantaburi Provinces) during the two periods, i.e., 1988-1993 (35 isolates) and 2003 (21 isolates). Finger prick blood samples (100-200 μl) were collected from patients with symptomatic and mono-infection with P. falciparum malaria and immediately mixed with 1 ml of transporting medium (RPMI 1640, 5% NaHCO3 and 10 units per miligram heparin; Thaithong et al. 1984). All isolates were continuously cultured using the methods of Trager and Jensen (Trager and Jensen 1976) with modifications. Laboratory strains of 3D7 (CQ sensitive: IC50 11 ± 1.5 nM) and K1 (CQ resistant: IC50 125 ± 15 nM) clones were used as positive control. The study protocol was reviewed and approved by the Ethics Committee of the Ministry of Public Health of Thailand.

In vitro drug susceptibility tests

Susceptibility of P. falciparum isolates to CQ, QN, and MQ were investigated based on the incorporation of [3H] hypoxanthine into parasite nucleic acids (Desjardins et al. 1979). Highly synchronous ring stage parasite was used in each assay. An aliquot of parasite inoculum (10 μl) with 1% parasitemia and 20% hematocrit was added into each well of microtiter plate. The final hematocrit for each well was 2%. The 96-well drug plates were dosed with antimalarial drugs at a total of eight final concentrations as follows: CQ (5, 10, 25, 50, 100, 150, 250, 500 nM), QN (10, 25, 50, 100, 150, 250, 500, 1,000 nM), and MQ (1, 5, 10, 25, 50, 100, 150, 200 nM). Each assay was performed in triplicate. IC50 values (drug concentrations that inhibit the uptake of [3H] hypoxanthine into parasite by 50%) were obtained from a log dose response curve plotted using the Grafit™ computer program (Erithacus Software Ltd., UK). The IC50 value was used to indicate antimalarial potency to allow a direct comparison of the activities of all the tested drugs.

Detection of pfcrt and pfmdr1 polymorphisms by PCR-RFLP

Genomic DNA was extracted using Chelex-resin (Biorad Co. Ltd., USA) according to the method of Wooden and colleague (Wooden et al. 1993). Previously published nested and PCR-RFLP methods were employed to detect pfmdr1 at the codons 86,184, 1034, 1042, and 1246 (Duraisingh and Cowman 2005; Duraisingh et al. 2000), and pfcrt mutations at the codons 76, 220, 271, 326, 356, and 371 (Fidock et al. 2000). The primers and reaction conditions used were according to the previously described methods (Duraisingh and Cowman 2005; Duraisingh et al. 2000; Fidock et al. 2000). PCR was performed in a total volume of 25 μl with the following reaction mixture: 0.1 μM of each primer, 2.5 mM MgCl2, 100 μM deoxynucleotides, 1× PCR buffer (100 mM KCl, 20 mM Tris-HCl pH 8.0), 2 μl of genomic DNA, and 0.5 unit of Taq DNA polymerase (Promega™). Genomic DNA of K1 and 3D7 clones were included in each experiment as positive (mutation) and negative (wild-type) control.

Detection of gene copy number of pfmdr1 by TaqMan real-time PCR

Pfmdr1 copy number was assessed by TaqMan real-time PCR (ABI sequence detector 7000; Applied Biosystems, Warrington, UK). The primers and a FAM-TAMRA (6-carboxyfluorescein-6-carboxytetramethyl rhodamine) probe specific to a conserved region of pfmdr1 and the primers and a VIC-TAMRA (VIC-6-carboxytetramethyl rhodamine) probe specific to ß-tubulin were designed with Prism 7000 sequence-detection software (Primer Express, Applied Biosystems) according to the method developed by Price (Price et al. 2004). Briefly, the amplification reactions was done as multiplex PCR in MicroAmp 96-well plates (Applied Biosystems) in a 25 µl reaction mixture containing TaqMan buffer pH 8.3 (8% glycerol, 0.625 U DNA polymerase, 5.5 mmol/l MgCl2, 300 μmol/l dNTP, 600 nmol/l passive reference dye 5-carboxy-X-rhodamine), 300 nmol/l of each forward and reverse primer, 100 nmol/l of each probe, and 5 µl of template DNA. The reaction was performed for 50 cycles (95°C for 15 s and at 58°C for 1 min). Fluorescence data was expressed as normalized reporter signal, calculated by dividing the amount of reporter signal by the passive reference signal. The detection threshold was set above the mean baseline value for fluorescence of the first 15 cycles. The threshold cycle (Ct) was when the increase in reporter signal was first detected above baseline. Results were analyzed by a comparative Ct method based on the assumption that the target (pfmdr1) and reference (ß tubulin) were amplified with the same efficiency within an appropriate range of DNA concentrations. The comparative ΔΔCt method was used: ΔΔCt = CtE – CtB, where CtE denotes the experimental C t and CtB the baseline Ct. Every TaqMan run contained two reference DNA samples from both 3D7 and K1 clones which contained only one copy numbers of pfmdr1. All reactions were performed in triplicate and results were rejected in case of non-exponential kinetics.

Statistical analysis

Distribution of the data was assessed using Kolmogorov-Smirnov test. Comparison of the prevalence of pfcrt and pfmdr1 gene mutations in P. falciparum isolates collected during the two time periods was performed using chi-square test. Association between pfmdr1 gene copy number and in vitro susceptibility of the parasites to CQ, QN, and MQ was done by student t test. Association between pfmdr1 copy number and allelic polymorphisms at codons 86, 184, 1034, 1042, and 1246 and in vitro susceptibility of P. falciparum isolates to CQ, QN, and MQ was assessed using Kruskal Wallis and Mann-Whitney U tests. The molecular markers influencing in vitro drug susceptibility level was determined by binary logistic regression. Statistical significance level was set at α = 0.05 for all tests.

Results

In vitro drug susceptibility of P. falciparum isolates

A total of 56 P. falciparum isolates collected from symptomatic malaria patients with mono-infections during the two periods, i.e., 1988-1993 (35 isolates) and 2003 (21 isolates), were investigated for their susceptibility to CQ, QN, and MQ. The criteria for discrimination between the resistant and sensitive isolates to CQ, QN, and MQ followed the previously described criteria (Cerutti et al. 1999; Pickard et al. 2003). CQ susceptibility was categorized into three levels, i.e., sensitive (S: IC50<25 nM), moderately resistant (MR: 25 ≤ IC50 < 100 nM), and highly resistant (HR: IC50 ≥ 100 nM). QN susceptibility was categorized into two levels, i.e., sensitive (S: IC50 < 500 nM), and resistant (R: IC50 ≥ 500 nM). MQ susceptibility was categorized into two levels, i.e., S (IC50 ≤ 24 nM), and R (IC50 > 24 nM). The range of IC50 values for CQ varied from 27 to 174 nM, with geometric mean (95% CI) value of 88 (79-99) nM. Twenty-eight (50%) isolates were classified as MR and HR. The IC50 values for QN ranged from 36 to 483 nM, with geometric mean (95%CI) value of 159 (135-187) nM. All isolates were categorized as QN-sensitive. The IC50 values for MQ ranged from 3 to 110 nM, with the mean (95% CI) values of 16 (13-21) nM. Thirty-six (64%), and 20 (36%) isolates were categorized as MQ-sensitive and MQ-resistant, respectively. The drug susceptibility profiles of P. falciparum isolates collected during the two periods to all drugs except MQ were similar. MQ susceptibility of parasite isolates collected during the period 1988-1993 was significantly higher (lower IC50) when compared with those collected during the year 2003. (Table 1)

Table 1 Susceptibility of P. falciparum isolates collected during the two periods (1988-1993 and 2003) to CQ, QN, and MQ
Full size table

Polymorphisms of pfmdr1 and pfcrt and pfmdr1 copy number

Pfcrt polymorphisms were commonly observed in the isolates studied. The distribution of patterns of pfmdr1 and pfcrt polymorphisms was shown in Fig. 1. The prevalence of the mutation at codons K76T, A220S, Q271E, N326S, and R371I were virtually 100%; only one isolate carried wild-type allele at codon 356 (I356). The prevalence of pfcrt polymorphism observed during both periods was similar, with TSESTI being the most prevalent haplotype.

Fig. 1
figure 1

The distribution of pfmdr1 and pfcrt haplotypes during the periods 1988-1993 and 2003

Full size image

The prevalence of pfmdr1 polymorphisms at codons N86Y, Y184F, S1034C, and N1042D were five (9%), 48 (86%), ten (18%), and 15 (27%), respectively. All isolates carried wild-type allele at codon 1246 (D1246). The pfmdr1 polymorphisms of the isolates collected during 1988-1993 were more diverse than that of 2003; 6 and 2 haplotypes of pfmdr1 were found in the isolates collected during 1988-1993 and 2003, respectively. The NFSND was the most prevalent haplotype found in both periods. A total of 19 isolates (34%) carried only one pfmdr1 gene copy, whereas 37 isolates (66%) carried more than one gene copies.

The associations between antimalarial drug susceptibility, genetic polymorphisms and gene amplification of P. falciparum

No association between pfcrt polymorphisms and the level of resistance of antimalarial drugs was found. All of the parasite isolates which were either CQ-resistant or CQ-sensitive carried mutant alleles at the codons 76, 220, 271, 326, and 371.

The association between antimalarial drug susceptibility of P. falciparum isolates, pfmdr1 polymorphisms and gene amplification was summarized in Table 2. The mutation of pfmdr1 at codon 1034 was found to be associated with the level of CQ resistance, whereas the mutations at the codons 184 and 1042 were associated with QN resistance. The mutations at the codons 1034 and 1042 improved susceptibility of the parasite to MQ, while the mutation at codon 184 led to decreased susceptibility to the drug. The parasite isolates containing more than one pfmdr1 gene copy exhibited decreased susceptibility to MQ comparing to those carrying only one gene copy.

Table 2 The association between pfmdr1 polymorphisms, gene copy number and susceptibility of the parasite isolates to CQ, QN, and MQ
Full size table

Pfcrt and pfmdr1 polymorphism/amplification patterns among parasite isolates collected during the two periods are summarized in Table 3. The mutation of pfcrt in P. falciparum isolates was stable with low prevalence during the observed 10-year period. The prevalence of N86Y, S1034C, and N1042D, but Y184F mutations were lower during the year 1988-2003 compared with 2003. Significant association between pfmdr1 polymorphisms and time periods was found only with the two codons, with the prevalence of the wild-type S1034C and N1042D polymorphisms significantly higher in isolates collected during 2003 compared with 1988-1993. In addition, isolates collected in 2003 with more than one pfmdr1 gene copy were found to be more prevalent compared with the period 1988-1993.

Table 3 Distribution of pfmdr1 polymorphisms of parasite isolates collected from different time periods
Full size table

Discussion

The present study investigated in vitro susceptibility profiles in association with patterns of molecular markers of resistance of P. falciparum isolates collected from the Thai-Cambodian border during the two periods (1988-1993 and 2003) to CQ, MQ, and QN. Our data showed no significant change in the pattern of mutations of pfcrt gene which is the key determinant of CQ resistance. The observation of high prevalence of mutations of pfcrt in almost all isolates at the codons A220S, Q271E, N326S, and R371I, was in line with the in vitro susceptibility data found in the present study as well as other previous studies (Chen et al. 2001, 2003; Fidock et al. 2000). The previous study reported virtually 100% mutation at codon T76 in all, except one isolate obtained from Chantaburi Province close to Thai-Cambodian border (Congpuong et al. 2005). This might suggest that, pfcrt polymorphism was sufficient to CQ-resistance but apart from pfcrt, other gene(s) may also play role in mediate level of CQ resistance.

Numerous in vitro studies have found the association between both copy number and polymorphism of the pfmdr1 gene and in vitro antimalarial drug susceptibility (Price et al. 1999; Chaiyaroj et al. 1999; Duraisingh and Cowman 2005; Pickard et al. 2003; Price et al. 2004). The current results support the role of pfmdr1 in modulating the degree of CQ, MQ, and QN susceptibility. Pfmdr1 SNPs have been proposed as mediating CQ resistance (corresponding to substitutions N86Y, Y184F, S1034C, N1042D, and D1246Y). In the present study, all these pfmdr1 polymorphisms were identified in P. falciparum isolates from Thailand collected during the period 1988-1993 and 2003. The majority of the isolates (91%, 83%, 74%, and 100%) carried wild-type (N86, S1034, N1042, D1246) at the codons 86, 1034, 1042, and 1246, respectively. The percentage of isolates carried mutant (Y184F) at codon 184 was 86%. All isolates were identified as wild-type (D1246) at codon 1246, which is in agreement with that previously reported in isolates from Thailand (Chaiyaroj et al. 1999; Congpuong et al. 2005; Price et al. 1999). These polymorphisms were found to be generally associated with reduced parasite susceptibility to CQ, QN, and MQ, except for the mutation at codons 1034 and 1042 which were associated with improved sensitivity to MQ. Gene amplification alone was found to be associated with reduced susceptibility of the parasite isolates to MQ. It has been suggested that the increase in pfmdr1 copy number might cause resistance to both artesunate and mefloquine (Price et al. 2004; Alker et al. 2007). In one study, pfmdr1 copy number was found to be associated with treatment failure following the combination regimen of mefloquine-artesunate but not with mefloquine-artemether (Price et al. 2006).

The prevalence of pfmdr1 polymorphisms including gene copy number of the wild types S1034C, N1042D were significantly different during the two observation periods. There was a trend of increasing prevalence of isolates with wild-type genotypes at these codons including that with more than one gene copy from the year 1988 towards 2003, and this appeared correlated with the decline in MQ susceptibility. In the absence of CQ pressure during the 10-year observation period, the prevalence of the dominant pattern of the wild-type pfmdr1 polymorphism NFSND (codon 86/184/1034/1042/1246) was significantly increased.

Our results showed that pfmdr1 polymorphisms influenced CQ susceptibility level, but did not resume the parasite susceptibility to sensitive level. Except for S1034C mutation, no association between pfmdr1 polymorphisms (N86Y, Y184F, N1042D, and D1246Y) and CQ susceptibility was found. All these results signify the possibility of the hypothesis that pfcrt mutation is the key determinant of CQ susceptibility but also requires involvement of the interactions with other genes. Conclusion on the association between pfmdr1 allele polymorphisms and CQ resistance had been contradictory among various studies, with either positive association (Basco et al. 1995), or lack of association (Adagu et al. 1995; Chaiyaroj et al. 1999; Wilson et al. 1993). The study by Reed et al. (2000) was later on, definitively confirmed its involvement in high level CQ resistance in such a way that mutation of pfmdr1 could modify CQ sensitivity only when present in a parasite expressing other resistance genes such as pfcrt (Reed et al. 2000). In addition, Sidhu et al. (2005) found that the introduction of pfmdr1 three mutations into the CQ-resistant 3BA6 line did not noticeably alter the degree of CQ resistance. These data suggest that the contribution of pfmdr1to CQ-resistant may only apply to a limited subset of genetic backgrounds (Sidhu et al. 2005).

The influence of pfmdr1 gene amplification on antimalarial susceptibility was observed with MQ in this study. Reduced susceptibility to MQ was found in the isolates carrying more than one pfmdr1 gene copies compared with those carrying only one copy. A few experiments selecting for resistance to MQ in vitro have demonstrated the influence of pfmdr1 gene amplification and over-expression on reduced sensitivity of the parasite to MQ (Sidhu et al. 2006).

Similar to CQ, susceptibility of the parasite to QN could also be influenced by multiple genes. We found stable pattern of QN susceptibility in parasite isolates collected from Thailand during the 10-year period. QN susceptibility does not seem to be linked with the pfmdr1 gene amplification in several studies of field isolates (Anderson et al. 2005; Pickard et al. 2003; Price et al. 2004). However, Sidhu et al (Sidhu et al. 2006) highlighted the importance of pfmdr1 copy number in determining P. falciparum susceptibility to QN using knockdown technique. Moreover, there was also evidence showing that QN sensitivity was modulated by other genes (Ferdig et al. 2004; Mu et al. 2003). Cooper et al (Cooper et al. 2007) selected parasite lines to become more resistant to QN and found that QN resistance was associated with pfcrt mutation that resulted in amino acid changes in PfCRT. Furthermore, SNPs in nine of 49 transporter gene sequences of P. falciparum were identified found to be related to QN susceptibility (Mu et al. 2003). However, there was no association between any of those genes and QN response in a cohort of samples from near the Thai-Myanmar border (Anderson et al. 2005)