Declining trend of Plasmodium falciparum dihydrofolate reductase (dhfr) and dihydropteroate synthase (dhps) mutant alleles after the withdrawal of Sulfadoxine-Pyrimethamine in North Western Ethiopia

Antimalarial drug resistance is one of the major challenges in global efforts of malaria control and elimination. In 1998, chloroquine was abandoned and replaced with sulfadoxine/pyrimethamine, which in turn was replaced with artemether/lumefantrine for the treatment of uncomplicated falciparum malaria in 2004. Sulfadoxine/pyrimethamine resistance is associated with mutations in dihydrofolate reductase (Pfdhfr) and dihydropteroate synthase (Pfdhps) genes. The prevalence of mutation in Pfdhfr and Pfdhps genes were evaluated and compared for a total of 159 isolates collected in two different time points, 2005 and 2007/08, from Pawe hospital, in North Western Ethiopia. The frequency of triple Pfdhfr mutation decreased significantly from 50.8% (32/63) to 15.9% (10/63) (P<0.001), while Pfdhps double mutation remained high and changed only marginally from 69.2% (45/65) to 55.4% (40/65) (P = 0.08). The combined Pfdhfr/Pfdhps quintuple mutation, which is strongly associated with sulfadoxine/pyrimethamine resistance, was significantly decreased from 40.7% (24/59) to 13.6% (8/59) (P<0.0001). On the whole, significant decline in mutant alleles and re-emergence of wild type alleles were observed. The change in the frequency is explained by the reduction of residual drug-resistant parasites caused by the strong drug pressure imposed when sulfadoxine/pyrimethamine was the first-line drug, followed by lower fitness of these resistant parasites in the absence of drug pressure. Despite the decrease in the frequency of mutant alleles, higher percentages of mutation remain prevalent in the study area in 2007/08 in both Pfdhfr and Pfdhps genes. Therefore, further multi-centered studies in different parts of the country will be required to assess the re-emergence of sulfadoxine/pyrimethamine sensitive parasites and to monitor and prevent the establishment of multi drug resistant parasites in this region.


Introduction
For more than 40 years, chloroquine was the first line anti-malarial drug in Ethiopia for the treatment of uncomplicated Plasmodium falciparum malaria. In 1998, chloroquine (CQ) was abandoned and replaced with sulfadoxine/pyrimethamine (SP). In 2004, SP was replaced with artemether/lumefantrine (AL) for the treatment of uncomplicated falciparum malaria [1,2]. SP is no longer recommended for the treatment of malaria but it remains a vital tool to reduce the burden of malaria in Africa. The World Health Organization (WHO) recommends SP for intermittent preventive treatment in pregnancy (IPTp), in infancy (IPTi) and in childhood (IPTc) [3][4][5]. In sub-Saharan Africa, 39 countries had adopted intermittent preventive treatment in pregnancy between 1993 and 2007 [6].
Following the withdrawal of CQ in 1993, Pfcrt mutation (K76T) gradually decreased and disappeared completely by 2001 in Malawi [19]. Similar findings have been reported in Tanzania [20], Kenya [21] and China [22]. This decline was correlated with the return of the clinical efficacy of CQ for the treatment of falciparum malaria in Malawi [23]. In Cambodia, however, alleles conferring CQ and SP resistance occur at a high frequency after the withdrawal of these drugs [24]. In Venezuela, the complete fixations of mutant Pfdhfr and Pfdhps alleles eight years after the withdrawal of SP were reported [25]. Zhou et al., 2009, reported decrease in the frequency of Pfdhfr and Pfdhps gene mutations after the withdrawal of SP in Peru. Decline in Pfdhfr and Pfdhps mutations after the withdrawal of SP has also been reported from Northern Ethiopia [26], Tanzania [27] and southern Mozambique [28].
In Ethiopia, high level of Pfdhfr triple mutations and Pfdhfr/Pfdhps quintuple mutations was reported from Jimma, Dilla and Bahirdar in 2004/05 [26,29,30]. A study conducted in Northern Ethiopia (Bahirdar) showed a significant decline in the triple Pfdhfr mutation from 78.6% in 2005 to 56.4% in 2008 and the quintuple mutations reduced from 60.6% to 37.2% between 2005 and 2008 [31]. In this study we evaluated change in the frequency of the mutant and wild type Pfdhfr and Pfdhps markers after the withdrawal of SP in 2004 in a perennial transmission setting. The objective of this study was to determine the change in the frequency of Pfdhfr and Pfdhps mutant and wild alleles in P. falciparum isolates collected in 2005 and 2007/08.

Study area
The study was conducted in the rural town of Pawe, in the North Western part of Ethiopia (Fig 1). The area is located at an altitude of 1050 meters above sea level with a mean annual temperature ranging from 16.2°C to 32.2°C, and the mean annual rainfall between 980 and 1200 mm occurring in two seasons from March to May and from June to December (Pawe Agricultural Research Centre). Pawe was one of the 14 sentinel sites, that were eco-epidemiologically selected, for drug resistance studies by the Federal Ministry of Health [2].

Study population
Plasmodium falciparum samples were collected from 80 patients in 2005 by the Ethiopian Public Health Institute (formerly known as Ethiopian Health and Nutrition Research Institute (EHNRI)). In 2007/08, we screened 602 patients attending Pawe hospital and 79 of them were positive for Plasmodium falciparum malaria by light microscopy. We included patients of all age, both sex and only those who lived in the study area for at least two years. The same inclusion criteria were used in 2005. Patients with severe malaria [32], severe malnutrition and serious underlying diseases according to the physicians comment were excluded from the study.
A finger-prick blood sample was taken for thick and thin blood smear. An experienced microscopist examined the Giemsa-stained blood smears. If the patient was infected with P. falciparum mono-infection and fulfilled the inclusion criteria, two to three drops of blood was collected on to a whatman filter paper (Krackeler Scientific Inc., New York). The filter paper was labeled, dried and stored at -20°C. After the blood sample was collected, additional thin and thick blood smears were made labeled and transported to EHNRI for re-confirmation by a different microscopist and parasite count. From the thick smears, parasite density was determined by counting the number of asexual parasites against 200 white blood cells. The thin smear was used for species determination. The national and Addis Ababa University (Department of Biology) ethical boards independently approved this study. A written informed consent was translated to the local languages of the study area and approved by the ethical review boards. During sample collection a written informed consent was obtained from each participant and parents or guardian and appropriate drug was given free of charge for patients according to the physician subscription.

DNA extraction and PCR amplification
Parasite genomic DNA was extracted from dried blood spot as previously described [33]. Pfdhfr and Pfdhps genes were amplified from 2-3μl of genomic DNA for the primary PCR and 2μL of the PCR product for the nested PCR. The sequence of the primers, PCR and cycling conditions are shown in S1 Table. Positive control genomic DNA from 3D7, Dd2, T994, HB3, T996, SL/D6 and IEC513/86 (S2 Table) and negative control were included in all PCR reactions. PCR products were visualized in agarose gel containing ethidium bromide. The bands were analyzed and the PCR products were used for the dot blot hybridization experiment. PCRs were repeated and optimized for the negative samples.

Preparation of dot blots
PCR products were chemically denatured, heated and then neutralized with equal volume ammonium acetate (pH = 7). The denatured PCR product was blotted by direct application onto Gene screen nylon membranes using dot blotter (Bio-Rad, UK). The membrane was incubated for 30 minutes and neutralized by washing in saline-sodium citrate buffer (2X SSC) for 60 seconds. The membrane was immersed in 0.4M Sodium Hydroxide for 30-60 seconds to ensure complete denaturation of immobilized DNA, and then rinsed in neutralizing buffer for 30 seconds. The cross-linking between the applied DNA and the membranes was done using UV cross-linker as previously described [34]. In each blot appropriate positive controls for Pfdhfr and Pfdhps (S2 Table) and negative control were included.

Labeling of oligonucleotide probes
Ten oligonucleotide probes were labeled with radioactive 32 P as described in previous studies (Table 1) [34,35]. Briefly, 1μl (10pMol) of each oligonucleotide probe (MWG Biotech, Germany) mixed with 1μl Ready-To-Go T4 Polynucleotide Kinase (Amersham Pharmacia Biotech, UK) and 47μl of DNase free water. 1μl of [γ-32P] ATP (Institutes of Isotopes, Hungary) was added and incubated for 30 minutes at 37°C. The reaction was stopped by 5μl of 250mM EDTA. The unincorporated [32p]-γ-ATP was removed by re-suspending the mixture in column according to the manufacturer guidelines (Amersham Pharmacia Biotech, UK). The purified sample was stored in -20°C until further use. Strict personal and environmental safety protocols of the national radiation regulation authority were followed.

Dot blot hybridization
The blot membrane was immersed in a hybridization buffer (Final concentration: 5X SSPE, 5X Denhardt's reagent, 0.5% SDS, 59.9% DNase-free water and 0.01 mg/ml sonicated salmon sperm) (Gibco BRL, UK) and incubated for 30 minutes at the specific hybridization temperature of each probe ( Table 1). The 32 P labeled oligonucleotide probe (1μl for every 1ml of the hybridization buffer) was added and hybridized overnight in rotary rocker hybridization oven. The membrane was washed once in 2X SSC and twice in 1X SSC/0.1% SDS with 20 and 10 minutes incubation, respectively. The washing solution was poured off and disposed according to local radiation regulations. The blot was wrapped in cling film and taped into an autoradiography cassette (Amersham Biosciences, France). It was exposed to X-ray film (Amersham Bioscience, France) at -70°C for 12-24 hours. The film was developed and results were scored.
Following autoradiography, probes were stripped off by two washes in 0.1M Sodium Hydroxide and rinsed in 5XSSC [34,35]. The blots were then re-hybridized with other probes as required.

Data analysis
The data was scored for each sample as mutant, wild and mixed (mutant and wild) based on the detection of the specific probes. The scored raw data is attached in S1 Dataset. We scored double, triple and quintuple mutations based on the detection of specific alleles in the isolates. For example, if we detect three Pfdhfr mutations in the same isolate, we score this isolate as a triple Pfdhfr mutant and the same for the Pfdhps and the quintuple mutation. The changes in the frequencies of mutations between groups were compared using the F-test. The data was analyzed using GraphPad Prism v6.0 (GraphPad Software Inc).  Table 2). The retrospective Table 1. List of oligonucleotide probes used for the detection of polymorphism in Pfdhfr and Pfdhps genes.

Characteristics of study participants
The data presented in this study is based on 159 samples tested for the change in the frequency of Pfdhfr and Pfdhps mutant and wild type alleles for the year 2005 (n = 80) and 2007/08 (n = 79). The study participants between the two study periods are comparable in terms of age, sex and parasite density ( Table 3) (Table 4).

Discussion
There are widely differing malaria endemicity and transmission regions in Ethiopia ranging from a sporadic cases in highland fringe areas to a perennial transmission south western lowlands [2]. Considering the substantial burden of malaria in this country, in 2004 multi-site survey demonstrated mean SP treatment failure rates of 36% and 72% within 14 and 28 days of follow-up, respectively [1]. Following this report, SP was replaced with artemisinin-based combination therapy, artemether/lumefantrine (AL). Different studies show that CQ withdrawal from use for a number of years has reversed resistance based on prevalence of Pfcrt resistance marker [19]. This was only possible because CQ was totally banned and it was not available to health facilities. Similarly, SP was banned in Ethiopia since 2004 and unlike many other African countries it has not been used for intermittent preventive treatment. This gives us the opportunity to study the temporal changes in the frequency of molecular markers associated with sulfadoxine/pyrimethamine resistance in the absence of drug pressure in a perennial malaria transmission setting. We compared 159 samples collected in two time points (2005 and 2007/8) by PCR based dot blot hybridization technique. We found a significant declining trend of Sulfadoxine/pyrimethamine mutation markers and re-emergence of wild parasites after the withdrawal of the drug in 2004. The quintuple mutation (Pfdhfr S108N/N51I/C59R/Pfdhps A437G/K540E), which was strongly correlated to SP treatment failure of P. falciparum malaria in south eastern Africa [13], was decreased significantly from 40.7% in 2005 to 13.6% in 2007/08. The level of mutation is much higher as compared to previous report of 1.4% prevalence in the absence of SP in the Ashanti Region of Ghana [36]. The quintuple predictor [13], Pfdhfr C59R and Pfdhps K540E double mutant, showed a statistically significant change from 51% in 2005 to 16.5% in 2007/08 (P<0.0001) supporting the hypothesis that the resistant strains are less fit in the absence of drug pressure.
Triple mutations in the Pfdhfr gene are associated with 60 to 70% rates of SP treatment failure [13]. The Pfdhfr triple mutations decreased from 51% to 16% (P<0.001) and the Pfdhps double mutation decreased from 69% and 55.4% (P = 0.08) indicating that Pfdhfr allele with triple mutations is being replaced faster than the Pfdhps allele with double mutations as the drug pressure was removed from the population, suggesting that pyrimethamine sensitivity is reemerging faster than sulfadoxine sensitivity. This is also supported by the increase in the triple wild Pfdhfr types from zero to 11.1% (P<0.0001) and Pfdhps double wild types from 13.8% to 30.8% (P = 0.006) in the two time points. These findings contrasts with the report from Peru, which showed a decrease in Pfdhps 437/540/581 mutation from 47% in 1997 to zero in 2005/6 while Pfdhfr triple mutation was decreased to 17% [37]. However, the transmission of malaria in this part of Ethiopia is higher than Peru and the time difference in the Peruvian study was five years. This indicates assessing the change in frequency of mutations in larger time points will give better resolution depending on the intensity of transmission. A similar study in southern Mozambique between 1999 and 2004 showed that the frequency of dhfr triple mutation increased from 1999 to 2004, however, the frequency of dhps double mutation increased to 2001 but declined to baseline levels by 2004. Quintuple mutations increased from 1999 to 2001 but decreased by 2004, which corresponded with replacement of SP with artemether-lumefantrine as malaria treatment policy in neighboring KwaZulu-Natal, South Africa [28]. Similarly, in Tanzania a community-based study compared the prevalence of mutation between IPTi intervention and comparison groups showed an increase in Pfdhfr triple mutation and stabilized frequencies of Pfdhps double mutant parasites following the withdrawal of SP as first-line treatment [27].
The slower decline in prevalence of mutations in Pfdhps (Figs 2 and 3) suggests that these mutations may be less deleterious to parasite fitness than are Pfdhfr mutations in this study area. Although the precise relationship between mutations in the Pfdhfr and Pfdhps genes in clinical sulfadoxine/pyrimethamine resistance is unclear, previous data showed that the presence of a sensitive Pfdhfr allele is highly predictive of sulfadoxine/pyrimethamine treatment success irrespective of the Pfdhps allele [38].
We found a significant decrease in all single mutations except Pfdhps A437G (from 75.4% to 69.8%, P = 0.35). These findings contradicts with a recent report from Bahir Dar, which is 230km away from this study area [26]. The prevalence of high level of single and combination mutations in Pfdhfr and Pfdhps alleles in 2005 is consistent with what was known about SP efficacy in this area during 2003. A mean treatment failure of 32.4% on the 14 days and 74.3% in the 28 days follow-up was reported before SP was withdrawn in 2003 [1]. Interestingly, SP-sensitive P. falciparum parasites have significantly reemerged after the withdrawal of the drug (Fig  3). Triple Pfdhfr wild alleles (Ser-108/Asn-51/Cys-59) increased from zero to 11.1%, which is in contrast with Ghana [36], a decrease from 13.3% in 2001 to 11.9% in 2003 in the absence of SP was reported. Double Pfdhps wild forms (Ala-437/Lys-540) increased from 13.8% to 30.8% and the most important quintuple Pfdhfr/Pfdhps wild forms increased from zero in 2005 to 10.2% in 2007/08, this finding is in accordance with previous report [23] and shows the reemergence SP sensitive parasites. These finding is also in contradiction with the reported changes in Bahir Dar [26].
The decrease in mutation and subsequent increase wild type parasites might be explained by the reduction of residual drug-resistant parasites, caused by the strong drug pressure imposed before 2004 when SP was the first-line drug, followed by lower fitness of these resistant parasites in the absence of drug pressure. Re-emergence of SP susceptible parasites support the hypothesis that drug-resistant P. falciparum parasite may be at competitive disadvantage when drug pressure is removed in agreement with epidemiological reports from northern Ethiopia [26], Tanzania [27], Southern Mozambique [28] and Peru for SP [37] and Malawi [19], Sudan [39] and Southeast Asia [40,41] for CQ. However, in Ghana, Cambodia and Venezuela SP resistance alleles have remained at a high frequency after the replacement of SP [24,25,36].
Although the change in the frequency of mutation is significant, the resistant alleles are still abundant in the study area which is explained by the following factors: the time difference between the two surveys was only two and half years and after the 2004 discontinuation of SP for the treatment of P. falciparum malaria, availability of AL was limited and 85% of the populations were living in rural areas with restricted access to health care giving rise to a high rate of presumptive treatment with available drugs like SP and CQ [42]. The other important factor is cross-resistance between SP and Trimethoprim/ sulfamethoxazole [43,44]. The uses of Trimethoprim/ sulfamethoxazole as prophylaxis against human immunodeficiency virus (HIV) associated opportunistic infections most likely make an essential contribution to SP resistance. In this study, we do not have the data for the HIV prevalence or positivity of the study populations and we cannot sort out the impact of trimethoprim/sulfamethoxazole in maintaining the Pfdhfr and Pfdhps mutation. Cross-resistance between SP and Trimethoprim/sulfamethoxazole appears to be a contributing factor rather than the exclusive factor responsible for less rapid decrease of resistant parasites [43].
The declining trend of mutant alleles and re-emergence of SP sensitive parasites in this perennial transmission area is encouraging despite the limitations of this study. First, this study did not include the change in the frequency of Pfdhps A581G mutation in this study area. However, different studies showed the association of A581G mutation and SP resistance in this part of Africa. Second, the methodologies used in this study detect the mutant and wild type alleles at each locus independent of the other. Sequencing of the full-length Pfdhfr and Pfdhps genes will give a better resolution to determine combinations of mutations. Finally, this study reports the change in frequency of mutation only with two time points, which are 2.5 years apart.
Large scale, multi-centered studies with multiple time points will be required to investigate the change in the frequency of mutation and re-emergence of SP sensitive parasites in different transmission areas of the country.
The recovery of SP-sensitive parasite populations in the study area after removal of drug pressure suggests an advantage of the native Pfdhfr and Pfdhps molecule over its mutant forms and points to the possible value of drug rotation strategies in antimalarial policies as we facing the advent of artemisinin resistance. The use of SP for the control and elimination of malaria together with the current tools will require much lower rates of mutation associated with SP resistance. Further multi-centered studies are required to monitor the change in the frequency of mutation, in order to monitor and prevent the establishment of multi drug resistant parasites in different parts of the country.
Supporting Information S1 Table. Primer Sequences and PCR conditions for the amplification reaction.