Plasmodium Falciparum Multidrug Resistance Gene-1 N86Y-Y184F-D1246Y Polymorphisms in Northern Nigeria: Implications for the Continued Use of Artemether-Lumefantrine in the Region

The analysis of single nucleotide polymorphism (SNPs) in drug-resistance associated genes is a commonly used strategy for the surveillance of antimalarial drug resistance in populations of parasites. The present study was designed and performed to provide genetic epidemiological data of the prevalence of N86Y-Y184F-D1246Y SNPs in Plasmodium falciparum multidrug resistance 1 (pfmdr1) in the malaria hotspot of Northern Nigeria.


Introduction
Antimalarial drug resistance is a major impediment to malaria chemotherapy in sub-Saharan Africa [9] largely because Plasmodium falciparum rapidly develops resistance to drugs [7]. Resistance to antimalarial drugs occurs through drug-selection of spontaneous mutations in P. falciparum that confer tolerance to the drug [54]. The selection and spread of drug resistant P. falciparum is facilitated by the rapid genome replication rate and by a relatively high mutation rate per generation of the parasite [8,10].
The speed of selection of mutants within parasite populations depends upon on the pharmacokinetics of the drug itself and its degree of usage within a given host population [9]. For many antimalarial drugs, molecular markers of parasite resistance are known. Surveillance of these markers in parasite populations can act as a proxy measure of the e cacy of drugs within that population, and can act as early warning signals of the emergence of resistance into new regions. Frequent and thorough molecular surveys of the prevalence of mutations associated with drug resistance can, therefore, inform regional drug policies.
Single nucleotide polymorphisms (SNPs) in the P. falciparum multidrug resistance gene (pfmdr1) and the P. falciparum chloroquine resistance transporter gene (pfcrt) are associated with parasite resistance to antimalarial drugs including Artimisinin based Combination Therapies (ACTs) and chloroquine [53]. African P. falciparum isolates may carry the resistant allele of pfcrt encoding the amino acids CVIET at codons 72-76 as well as a variety of polymorphic pfmdr1 alleles which have originated and spread within the African continent [3,33,40]. The pfmdr1 gene is a structural homologue of the mammalian multidrug resistance gene encoding a P-glycoprotein homologue-1 (Pgh1) multi-drug resistant transporter [17] and is expressed into a PfMDR1 transporter located in the P. falciparum food vacuole.
Mutations in PfMDR1 are associated with enhanced e ux of diverse antimalarial drugs reducing their intracellular accumulation [17,42]. Single nucleotide polymorphisms (SNPs) in pfmdr1 are associated with resistance to multiple antimalarial drugs [12,44]. Several codons in pfmdr1 have been putatively linked with changes in the parasite's sensitivity to antimalarial drugs, but codons N86Y, Y184F and D1246Y are uniquely associated with changes in artemether-lumefantrine (AL) and artesunateamodiaquine (AS-AQ) e cacies in sub-Saharan Africa [14]. While the pfmdr1 86Y allele was strongly associated with chloroquine (CQ) and amodiaquine (AQ) resistance [16,49], 1246Y alleles were shown to confer resistance to quinine (QN) and possess the capacity to increase parasite susceptibility to me oquine (MQ), halofantrine (HF) and artemisinin (ART) [20,24].
The mutant pfmdr1 86Y and 1246Y alleles have also been linked to reduced sensitivity to AS-AQ, whereas the wild-type pfmdr1 N86 and D1246 alleles are linked to resistance against AL [34,41]. In Africa, the common use of AL and AS-AQ in the treatment of uncomplicated malaria has been linked with the emergence of pfmdr1 N86Y, Y184F and D1246Y SNPs [31], and the prevalence of these mutations are frequently used for evaluating AL and AS-AQ e cacies [36]. Several studies have shown that parasites carrying a combination of pfmdr1 N86, 184F, and D1246 (the "NFD" haplotype) display decreased susceptibility to AL and that treatment with AL can select for such a haplotype [4,46].
Nigeria accounts for 25% of global cases of malaria and an estimated 50% of the country's population suffer at least one episode of malaria every year while under-ve children experience an average of 2-4 attacks in a year [55]. Plasmodium falciparum is stably and perennially transmitted in all parts of the country [39], with transmission increased during the wet season compared to the dry [18,43].
North-West and North-East Nigeria have so far been identi ed as hotspots of malaria in relation to the southern parts of the country due primarily to climatic and environmental conditions [37]. However, the North-West region of the country suffers a much higher P. falciparum transmission rate than the other regions including North-East Nigeria [28].
The frontline drug for malaria chemotherapy in the country was chloroquine until 2005 when it was withdrawn as a result of resistance. Subsequently, ACTs especially artemether-lumefantrine (AL) were recommended in all parts of the country as the frontline chemotherapy for uncomplicated malaria.
Unfortunately, several reports investigating molecular markers of antimalarial resistance have suggested an increased risk of parasite tolerance to AL [13,20,[30][31]. However, there is no baseline data involving pfmdr1 SNPs in both North-West and North-East Nigeria since the withdrawal of CQ and adoption of AL in Northern Nigeria. In this study, the distributions of the pfmdr1 N86Y, Y184F and D1246Y SNPs across the North-West and-East Nigeria were investigated.

Description of study sites
Nigeria's North-West and North-East are two out of the six geo-political zones of Nigeria. The North-West is made up of seven states and is home to a population of over 35 million people whilst the North-East comprises six states with a population of over 18 million [52]. Two states from the North-West; Kano (longitude 7° 10´E, 10° 35´E and latitude 10° 25´N, 13° 53´N) and Kaduna (longitudes 7° 23´E and 7° 29´E and latitudes10° 25´N and 10° 36´N) with a combined population of 15,450,244 were randomly selected for inclusion in this study while Yobe (longitude13.5° E and latitude 11° N) and Adamawa (longitude 11°a nd 14°E and latitude 7° and 11°N ) states with a combined population of 5,489,692 were similarly selected from the North-Eastern region [52]. Other relevant details about the study sites are indicated in Fig. 1.

Selection Criteria
In this study, patients who presented with symptoms of uncomplicated malaria across all ages and did not take any antimalarial drug prior to arrival to the facilities were included whilst those who presented with severe malaria were excluded.

Sample Collection
Between June and November 2017, thick and thin lm microscopy was used to con rm P. falciparum positivity of malaria symptomatic patients that attended selected health facilities within the study sites. The total number of samples collected from Kano, Kaduna, Adamawa and Yobe were 250, 150, 150 and 200 respectively. 10 µL of microscopically con rmed P. falciparum parasitized blood samples were spotted on four different positions onto Whatman-3MM lter papers and allowed to dry at room temperature. Each sample was placed in sachets containing desiccant, and was preserved in a refrigerator at 4 °C.
Genomic DNA isolation, ampli cation and genotyping of pfmdr1 Three discs (3 mm/disc) were punched from the P. falciparum-positive dried blood spots and the punch sterilized between punches. The discs were used to extract genomic DNA using a QIAamp DNA Mini Kit (Qiagen Inc, Japan) according to the manufacturer's instructions. Nested polymerase chain reaction (PCR) was carried out as described by Humphreys et al. [20] with slight modi cations to the PCR cycle programs ( Table 1). The nested PCR runs for the ampli cation of two separate pfmdr1 segments S1 and S2, spanning codons 86-184, and − 1246 respectively were performed with a 10 µL nal master mixture and 200 ηM primers, and 1U ExTaq polymerase (Takara-Bio, Japan). Two µL of isolated genomic DNA was added to each of the rst PCR master mixtures and run in the thermocycler. At the completion of the rst PCR runs, 1 µL of each of the resulting ampli ed fragments were further used as templates in 10 µL of secondary PCR reactions. The PCR reactions were run along with parasite free genomic DNA (negative) and 3D7 clone P. falciparum genomic DNA (positive) controls. Two µL of each PCR product was evaluated by electrophoresis on 1.5% agarose gels that were stained with Midori Green Advance and visualized under ultraviolet light. The remaining 8 µL nested PCR products were stored at -30 °C. All nested amplicons were subsequently puri ed using the one step ExoSAP-IT (ThermoFisher Scienti c, Japan) puri cation kit and the resulting products subjected to BigDye Terminator (v3.1) Cycle Sequencing (ThermoFisher Scienti c, Japan). Sequences were analyzed using BioEdit Sequence Alignment Editor (v7.0.5.3) while pfmdr1 SNPs were determined using MEGA5 software (Build#:5130611) in reference to the pfmdr1 sequence of P. falciparum deposited at the NCBI database [Accession Number X56851]. Consequently, the prevalences of wild and mutant pfmdr1 across the three codons were calculated using the following formula: Each PCR run was preceded by an initial denaturation at 94 °C for 3minutes Pfmdr1 N86Y-Y184F-D1246Y Haplotypes The pfmdr1 haplotypes used in this study were based on eight previously reported haplotypes associated with artemether lumefantrine-tolerance; pfmdr1 N86Y, Y184F and D1246Y single nucleotide polymorphisms in different P. falciparum populations from Africa [22,27].

Statistical analysis
All data were statistically analyzed using Pearson Chi-square and Fisher's exact (FE) tests of Graph-Pad Prism (8.1.0) and P values 0.05 were considered to be statistically signi cant.

Results
Prevalence of SNPs in pfmdr1 codons 86, 186 and 1246 Of the 750 P. falciparum positive samples collected from the four states in northern Nigeria, 500 were successfully genotyped for the pfmdr1 N86Y, Y184F and D1246Y alleles. Six of the genotyped pfmdr1 sequences were deposited in the GenBank with accession numbers MT472640, MT472641, MT472642, MT495456, MT495458, and MT495459 on the basis of presence or absence of mutations in the three codons of the gene. The prevalence of the mutant pfmdr1 86Y allele was observed to be signi cantly ( ² = 10.47, P < 0.05) different across the states, with highest prevalence of 12.5% obtained in Kaduna state, followed by Kano with 4.68% and 2% in Adamawa state ( Fig. 2A) (Fig. 2C).
The regional distributions of the mutant pfmdr1 86Y, 184F and 1246Y alleles are shown in Fig. 3. Based on the results, an overall regional prevalence of 7.49% and 3% for the pfmdr1 86Y allele was recorded in North-West and North-East Nigeria, respectively. However, the observed difference was not signi cant ( ² = 1.68, P > 0.05), Fig. 3A. Similarly, the prevalence of the pfmdr1184F and 1246Y mutants were not signi cantly different between the North-West and North-East Nigeria, whereas the prevalence of the 184F allele differed signi cantly between these two regions; 68.91% and 56.22%, in the North-West and North-East respectively ( ² = 3.60, P > 0.05) (Fig. 3B). The prevalence of the pfmdr1 1246Y allele in the North-West and North-East Nigeria was 1.93% and 3% respectively ( ² = 0.21, P > 0.05) (Fig. 3C).

Analysis of pfmdr1 Haplotypes
The distribution of pfmdr1 haplotypes in the four Northern Nigerian states is shown in Fig. 4A. The pfmdr1 N86Y, Y184F and D1246Y mutations were constructed into NYD, NYY, NFY, NFD, YYY, YYD, YFD and YFY haplotypes. Out of the 500 pfmdr1 samples genotyped, a total of 492 haplotypes were constructed (sub-divided into seven different types). As shown in the Fig. 3, there was a signi cant difference in the prevalences of all pfmdr1 haplotypes across the locations ( ² = 36.05, P < 0.05); the NFD pfmdr1 haplotype was highest in Kano state with a prevalence of 69.81%, the NYD haplotype was highest in Adamawa with a prevalence of 49%, whilst the pfmdr1 YFD haplotype predominated in Kaduna with a prevalence of 11.46%. The pfmdr1 YYY haplotype was not detected (Fig. 4A). Figure 4B showed the distribution of the pfmdr1 haplotypes across North-West and North-East Nigeria, but in contrast to the states distribution, no signi cant difference was observed ( ² = 4.26, P > 0.05). The pfmdr1 NFD haplotype was highest in the North-West with a prevalence of 61.96%, the NYD haplotype was highest in the North-East with a prevalence of 41.63%, and the YFD haplotype was highest in the North-West with a prevalence of 5.88%.

Discussion
There has been recent report of AL failure in Northern Nigeria [5]. Mutations in several genes, including pfmdr1, pfcrt and pfk13 are associated with variation in parasite sensitivity to a range of drugs [6,15,26]. The pfmdr1 mutations N86Y, Y184F and D1246Y SNPs are thought to modulate susceptibility to CQ, AL and AS-AQ [53]. We found that, pfmdr1 184F and 86Y alleles predominated in North-West Nigeria while 1246Y was higher in the North-East.
Alleles of pfmdr1 carrying the wild type N86 residue are associated with higher IC 50 and IC 90 values for LMF, MFQ and DHA, while the alternative 86Y residue seems to confer increased resistance against CQ and AQ [53]. Similarly, there are varying epidemiological reports on the prevalence and consequences of pfmdr1 N86Y polymorphisms from different parts of the world. For example, Ibraheem et al. [21] reported that pfmdr1 mutations are geographically con ned and have inconsistent distributions from one geographic region to another.
Following adoption of ACTs in many African countries, some studies from West Africa have linked the prevalence of the pfmdr1 N86 allele to selection by AL [36]. Therefore, the high prevalence of pfmdr1 N86 allele observed in the present study might be suggestive of possible AL pressure in all the states. In addition, the nding might also suggest that the e cacy of the LMF component of ACTs is susceptible to the emergence of decreased tolerance in the local P. falciparum populations, as the presence of pfmdr1 N86 is critical in the initiation of resistance to LMF in vivo and that its selection primarily follows reinfection and recrudescence events associated with the elimination stage of LMF, 4-5 days after artemether clearance [45].
Some reports have associated a rise in the prevalence of pfmdr1 86Y alleles with increasing CQ resistance [11][12]25]. The low prevalence of pfmdr1 86Y in Adamawa and Yobe raises the possibility that CQ may be effective against P. falciparum malaria in North-Eastern Nigeria once again, although this would be presumably tempered by CQ-resistance associated mutations in pfcrt, which we have not assayed here. It is possible that the selection of pfmdr1 86Y allele in this region was aided by the cessation of CQ usage due to the emergence of resistance. The high prevalence of this mutation across Northern Nigeria may indicate that the e cacy of AL is at risk in this region, but raises the possibility that CQ may be effective in the chemotherapy of uncomplicated malaria here.
The effect of pfmdr1 Y184F polymorphisms on the e cacy of antimalarial drugs has been shown by various in vitro studies to be insigni cant [15,53]. Variations in the IC 50 of antimalarial drugs between parasite lines expressing wild type pfmdr1-Y184 or mutant 184F were shown to be closely linked to either of pfmdr1 N86 or 86Y alleles and not the 184F allele [53]. However, several epidemiological studies on the prevalence of pfmdr1 Y184F polymorphisms have shown that the Y184 allele is predominantly con ned to East and Central Africa while the mutant 184F allele predominates in West Africa [1,32,36,48]. Indeed, reports of the high occurrence of the mutant pfmdr1 184F in West Africa were corroborated by the present ndings in which we show that the prevalence of pfmdr1 184F was high in all the states, and especially in Kano, and that its prevalence is higher in North-West compared to North-East Nigeria. This mutation has been previously linked to a reduction in susceptibility to LMF and /or ART [19]. It is perhaps unsurprising, given this, that we nd a relatively high prevalence of this mutation in regions where AL is rst line intervention against uncomplicated malaria.
Despite the fact that most in vitro and in vivo studies have not strongly associated changes in mutant pfmdr1 184F with alterations in antimalarial drug response [47], the high prevalence of mutant pfmdr1 184F obtained in the present study might indicate an increased propensity towards AL treatment. In agreement with this is the fact that there is a lower prevalence of mutant pfmdr1 184F alleles in those East African countries that use AS-AQ or AL/AS-AQ as rst line treatments compared to West African countries strictly managing uncomplicated malaria with AL [36]. The observed high prevalence of pfmdr1 184F across the region suggests that AS-AQ or DHAP should be used as an alternative to AL as rst line treatments for uncomplicated falciparum malaria here.
The pfmdr1 D1246Y mutation affects P. falciparum susceptibility to various antimalarials including QN, MFQ, (HF), CQ and ART, with the latter two drugs affected in a strain speci c manner [14,42]. The observed low prevalence of mutant pfmdr1 1246Y alleles compared to the wild type in this study is consistent with reports from Southern Nigeria [38] as well as other West and East African countries that adopted AL as a front-line antimalarial therapy for uncomplicated malaria [1,36]. Countries in Central Africa have observed an unsteady increase in the prevalence of the pfmdr1 D1246 allele, possibly due to the selective pressure of AS-AQ [2,6].
Several reports from Africa have suggested that linkage between pfmdr1 N86Y/Y184F/D1246Y results in haplotypes with particular phenotypic characteristics that may be selected depending on the particular drugs that the population is exposed to [50][51]. The occurrence of pfmdr1 NFD and NYD haplotypes, for example, may result from AL selection while the pfmdr1 YYY haplotype may be favoured in regions where parasites are exposed to AS-AQ, DHAP and CQ [15,32]. The treatment of uncomplicated malaria with AL often selects pfmdr1 haplotypes bearing the N86 allele [29,48]. We found a predominance of the pfmdr1 NFD haplotype in Northern Nigeria with Kano, and a complete absence of the pfmdr1 YYY haplotype. These ndings are line with the selective effect of AL on the NFD haplotype, and may indicate a loss of susceptibility to AL treatment by parasites in this region.

Conclusions
In conclusion, Kaduna and Kano States had higher prevalences of the pfmdr1 86Y allele than Yobe and Adamawa indicating differential selection in North-East and North-West Nigeria, possibly due to differing population density of these regions. Furthermore, there was a very high prevalence of pfmdr1 NFD and NYD haplotypes which suggests that AL e cacy may be reduced in both regions. Overall, the scarcity of pfmdr1 YYY with a high prevalence of NFD haplotypes could inform a rational antimalarial drug policy shift from AL to either AS-AQ or DHAP and CQ in this region.