Polymorphisms in Plasmodium falciparum dihydropteroate synthetase and dihydrofolate reductase genes in Nigerian children with uncomplicated malaria using high-resolution melting technique

In 2005, the Nigerian Federal Ministry of Health revised the treatment policy for uncomplicated malaria with the introduction of artemisinin-based combination therapies (ACTs). This policy change discouraged the use of Sulphadoxine-pyrimethamine (SP) as the second-line treatment of uncomplicated falciparum malaria. However, SP is used as an intermittent preventive treatment of malaria in pregnancy (IPTp) and seasonal malaria chemoprevention (SMC) in children aged 3–59 months. There have been increasing reports of SP resistance especially in the non-pregnant population in Nigeria, thus, the need to continually monitor the efficacy of SP as IPTp and SMC by estimating polymorphisms in dihydropteroate synthetase (dhps) and dihydrofolate reductase (dhfr) genes associated with SP resistance. The high resolution-melting (HRM) assay was used to investigate polymorphisms in codons 51, 59, 108 and 164 of the dhfr gene and codons 437, 540, 581 and 613 of the dhps gene. DNA was extracted from 271 dried bloodspot filter paper samples obtained from children (< 5 years old) with uncomplicated malaria. The dhfr triple mutant I51R59N108, dhps double mutant G437G581 and quadruple dhfr I51R59N108 + dhps G437 mutant haplotypes were observed in 80.8%, 13.7% and 52.8% parasites, respectively. Although the quintuple dhfr I51R59N108 + dhps G437E540 and sextuple dhfr I51R59N108 + dhps G437E540G581 mutant haplotypes linked with in-vivo and in-vitro SP resistance were not detected, constant surveillance of these haplotypes should be done in the country to detect any change in prevalence.

Merozoite surface protein genotyping of Plasmodium falciparum. The family-specific polymorphic length markers of msp-2 and msp-1 were used for genotyping the P. falciparum in children considered for this study. In general, 180 (66.8%) children had polyclonal infections ( Table 2). The mean complexity of infection (mCOI) for parasites across all population was 2.3. Allelic family distributions for both msp-2 and msp-1 per State is represented in Table 2. Overall, the 3D7 allelic family was the most amplified (50.7%) in the msp-2 polymorphic marker while the K1 allelic family was the most detected (18.4%) msp-1 polymorphic marker.
Polymorphisms in the dhfr gene. From the 271 children considered, 241 (88.9%), 245 (90.4%) and 219 (80.8%) were infected with parasites that harboured the mutant I 51 , R 59 and N 108 alleles respectively. None of the children was infected with parasites habouring the mutant L 164 allele. Distribution of the mutant I 51 , R 59 and N 108 alleles were significantly higher in polyclonal infections than in monoclonal infections (p < 0.05 for each allele). All isolates obtained from children enrolled in Imo State were infected with parasites that harboured only the mutant I 51 and R 59 alleles (Fig. 1a,b).   www.nature.com/scientificreports/ Double mutant R 59 N 108 haplotype was only recorded in two children from Kwara and Sokoto States (Northern States) (Fig. 2). Parasites harbouring the dhps double mutant G 437 G 581 , single mutant G 437 and single mutant G 581 haplotypes were observed in 13.7%, 66.1% and 2.6% respectively of the 271 children considered (Fig. 3). There was no triple and quadruple dhps mutant haplotype because none of the isolates harboured mutant allele of dhps 540. There was no difference in the distribution of the double mutant G 437 G 581 and single mutant G 437 haplotypes in the Northern and Southern States (p > 0.05). www.nature.com/scientificreports/ were considered (Table 3). Although most of the dhfr I 51 R 59 N 108 triple mutant haplotype was observed in polyclonal infections as only eight were observed in monoclonal infections, there was no significant difference in distribution based on clonality (p > 0.05) or location, i.e., Northern and Southern States (p > 0.05). The quadruple mutant haplotype was observed in 143 (52.8%) of the 271 children (Table 3). Eighty-three (58.0%) of the 143 were observed in polyclonal infections but there was no significant difference in the distribution of this haplotype based on clonality (polyclonal vs. monoclonal) (p > 0.05) or location, i.e., Northern and Southern States (p > 0.05). None of the 271 children was infected with parasites harbouring the quintuple mutation and sextuple mutation as the E 540 mutant allele was absent (Table 3).

Discussion
This study assessed the status of circulating dhfr and dhps haplotypes by describing polymorphisms on codons 51, 59, 108 and 164 of dhfr gene and codons 437, 540, 581 and 613 of dhps gene and estimated the prevalence of dhfr + dhps combined mutant haplotypes in 271 parasites obtained from children (< 5 years) children with uncomplicated falciparum malaria in Nigeria 10 years after treatment policy was changed to ACTs. Polymorphism data from our study showed high prevalence of mutant I 51 (88.9%) and N 108 (80.8%) dhfr alleles and mutant G 437 (82.3%) dhps allele. Similar prevalence of these mutant dhfr and dhps alleles have been recorded in Nigeria 11,13 and other West African countries 14,15 . Although prevalence of these mutant alleles are generally high in West Africa 16 , lower prevalence (26.5-56.25) have been recorded in other West African countries 4,17 . The exact reason for the difference in prevalence amongst these West African countries may be as a result of the varied use of SP in these countries 18 . Also, P. falciparum and other disease etiologies exist as co-infections in patients in these areas. It is equally plausible that the use of other sulpha-related drugs in the treatment of these co-infections may select for these mutations in the P. falciparum genome at varying levels 19 .
Sulfadoxine-pyrimethamine was previously used as weekly prophylaxis for malaria during which the mutant E 540 allele was recorded 2,5 . However, our study which was conducted when the therapeutic use of SP had changed from weekly prophylaxis to IPTp and SMCs, showed the absence of the mutant E 540 allele. Similar trend has been observed in recent studies in Nigeria 11,19 . The supposed disappearance of this mutant allele is perhaps, as a result of the reduced SP drug pressure in the country due to this treatment policy change.
Pearce et al. 20 stated the importance of measuring the frequency of haplotypes as against the prevalence of each point mutation separately as haplotypes are determinants of resistance levels. We observed a high frequency (80.8%) of the dhfr triple mutant haplotype (I 51 R 59 N 108 ) which suggests the persistent circulation of similar parasites as those reported in earlier studies post-ACT introduction 11,12,18 . These parasites are probably selected for as a result of the SP drug pressure, as this drug was not completely withdrawn in the country but rather used as IPTp and SMCs till date. The use of drugs such trimethoprin sulfamethazole targeting dhfr genes in Pneumocystis carnii in an environment where malaria and HIV coinfections is common, could also be responsible for the selection of this haplotype in Plasmodium falciparum populations in Nigeria 5 . The occurrence of this mutant haplotype at such a high frequency is worrisome as such triple mutations in the dhfr gene has been associated with a 1.5-to threefold higher pyrimethamine resistance in vitro than I 51 N 108 or R 59 N 108 double mutations 5 . Thus, the efficacy of pyrimethamine as a partner drug in SP's use as IPTp and SMCs is threatened. Although the double dhps mutant haplotype (G 437 G 581 ) was observed in our study (12.9%), the absence of the E 540 mutation in the dhps gene in combination with either the single mutant G 437 or double mutant G 437 G 581 in this study is desired as the double mutant G 437 E 540 haplotype is essential for sulfadoxine resistance 5 .
We also observed high levels of polyclonal infections in this study as most of the States considered had > 60% proportion of children with polyclonal infections. Further analysis of data revealed that the mutant dhfr (I 51 , R 59 and N 108 ) and mutant dhps (G 437 and G 581 ) alleles were significantly higher in polyclonal infections than in monoclonal infections (p < 0.05 for each mutant allele). Also, the dhfr triple mutant I 51 R 59 N 108 haplotype was significantly higher in polyclonal infections (p < 0.05). These observations may be problematic as high levels of polyclonality is linked to increased parasite transmission and diversity. This may result in the increase in spread of these mutant alleles and haplotype within the country. This can jeopardise the use of SPs as IPTp and SMCs in Nigeria.
Unpublished P. falciparum microsatellite data 21 confirmed the high intra-population diversity observed using the msp-1 and msp-2 polymorphic genes but also revealed low population differentiation in these parasites from the five parasite populations (Nigeria States). This suggests that, despite the high parasite diversity observed, parasites were genetically similar across the country. This may be responsible for the observed similarities in the distribution of the combined dhfr + dhps triple mutant haplotype (dhfr I 51 R 59 N 108 : p > 0.05) and quadruple mutant haplotype (dhfr I 51 R 59 N 108 + dhps G 437 : p > 0.05) in both Northern and Southern States of the country considered in this study. www.nature.com/scientificreports/ The quintuple mutant haplotype (dhfr I 51 R 59 N 108 + dhps G 437 E 540 ) that was earlier reported (2005) in Nigeria 5 was absent in this study. This is possibly due to the reduced SP drug pressure as a result of policy change from SP as second-line treatment of malaria to ACTs. Both the quintuple and sextuple mutant haplotypes have been strongly linked with in vivo and in vitro SP resistance 22 in Southern and East Africa 23,24 and their absence in this current study, is not only beneficial, but consistent with reports from other West African countries 4,25 . Nevertheless, continuous monitoring for re-emergence of dhfr I 51 R 59 N 108 + dhps G 437 E 540 and emergence of dhfr I 51 R 59 N 108 + dhps G 437 E 540 G 581 haplotypes should be maintained in the country to detect any change in the recorded prevalence. This would ensure that alternative control measures are rapidly put in place to prevent the spread of these haplotypes within the country, which if not checked, will lead to reduced efficacy of SP as IPTp and SMCs.

Methods
Study site. This study is part of a national drug therapeutic efficacy testing (DTET) study for monitoring antimalarial efficacies of artemether-lumefantrine (AL), artesunate-amodiaquine (AA) and dihydroartemisininpiperaquine (DHP) in the treatment of acute uncomplicated falciparum infections in children under the age of five years. Samples considered for analysis in this study were obtained from five Nigerian States which were sub-classified into Northern region (Adamawa, Sokoto and Kwara States) and Southern region (Bayelsa and Imo States) (Fig. 4) Patients enrolment criteria and sample collection. Description of patient enrolment at sentinel locations was initially discussed in an earlier study 26 . Two to three drops of finger-prick blood were blotted on 3 mm Whatman filter paper (Whatman International Limited, Maidstone, United Kingdom) before treatment (Day 0) and during follow-up on days 1-3, 7, 14, 21, 28, 35 and 42 post-treatment. Blood samples impregnated on filter papers were allowed to air-dry appropriately at room temperature and stored in airtight envelopes with silica gels. DNA extraction. DNA was extracted from dry blood spot (DBS), i.e., blood impregnated filter papers of day 0 samples (before treatment) for the detection of polymorphisms of the drug resistance markers and to www.nature.com/scientificreports/ determine parasite genetic diversity as previously described 27 . A Qiagen DNA extraction kit (Qiagen, Hilden, Germany) was used to extract DNA from DBS, following the manufacturer's protocol. Briefly, a quarter of the DBS was used for extraction, and DNA content was eluted in a final volume of 60 μl with buffer AE.
Genotyping Plasmodium falciparum using the msp-1 and msp-2 gene. The polymorphic length markers msp-2 and msp-1 were amplified by nested PCR as previously described 28 . The Glurp polymorphic marker was not considered in this study due to the low PCR amplification reported in Nigeria 29 . PCR amplification was performed on a thermocycler (Eppendorf Vapo. Protect Mastercycler pro, Germany) in a final volume of 25 μl. Two per cent (2%) agarose gel was used for the resolution of PCR amplicons. The amplified products were sized against a 100-base pair (bp) DNA molecular weight marker (New England Biolabs, Beverly, MA) and visualised using a gel visualisation box (Syngene, UK). Interpretations were made based on the number of parasite clones present in a sample. Briefly, infections were defined as polyclonal if parasites from a single patient showed more than one allelic family or more than one amplicon fragment in a single allelic family of the gene.
Infections were defined as monoclonal if an isolate had a single amplicon fragment in one allelic family and the other allelic family(ies) was (were) not amplified 6 .
High resolution melting drug resistance assay. High resolution melting (HRM) assay was performed as previously described 30 . Briefly, the 10X primer-probe mix and reaction mix was prepared. One microliter of the quantified DNA sample was dispensed in PCR well containing 9.0 μl reaction mix. PCR cycling and melting conditions used were those described earlier 30