Molecular characterization of Plasmodium falciparum antifolate resistance markers in Thailand between 2008 and 2016

Background Resistance to anti-malarials is a major threat to the control and elimination of malaria. Sulfadoxine–pyrimethamine (SP) anti-malarial treatment was used as a national policy for treatment of uncomplicated falciparum malaria in Thailand from 1973 to 1990. In order to determine whether withdrawal of this antifolate drug has led to restoration of SP sensitivity, the prevalence of genetic markers of SP resistance was assessed in historical Thai samples. Methods Plasmodium falciparum DNA was collected from the Thailand–Myanmar, Thailand–Malaysia and Thailand–Cambodia borders during 2008–2016 (N = 233). Semi-nested PCR and nucleotide sequencing were used to assess mutations in Plasmodium falciparum dihydrofolate reductase (pfdhfr), P. falciparum dihydropteroate synthase (pfdhps). Gene amplification of Plasmodium falcipaurm GTP cyclohydrolase-1 (pfgch1) was assessed by quantitative real-time PCR. The association between pfdhfr/pfdhps mutations and pfgch1 copy numbers were evaluated. Results Mutations in pfdhfr/pfdhsp and pfgch1 copy number fluctuated overtime through the study period. Altogether, 14 unique pfdhfr–pdfhps haplotypes collectively containing quadruple to octuple mutations were identified. High variation in pfdhfr–pfdhps haplotypes and a high proportion of pfgch1 multiple copy number (51% (73/146)) were observed on the Thailand–Myanmar border compared to other parts of Thailand. Overall, the prevalence of septuple mutations was observed for pfdhfr–pfdhps haplotypes. In particular, the prevalence of pfdhfr–pfdhps, septuple mutation was observed in the Thailand–Myanmar (50%, 73/146) and Thailand–Cambodia (65%, 26/40) border. In Thailand–Malaysia border, majority of the pfdhfr–pfdhps haplotypes transaction from quadruple (90%, 9/10) to quintuple (65%, 24/37) during 2008–2016. Within the pfdhfr–pfdhps haplotypes, during 2008–2013 the pfdhps A/S436F mutation was observed only in Thailand–Myanmar border (9%, 10/107), while it was not identified later. In general, significant correlation was observed between the prevalence of pfdhfr I164L (ϕ = 0.213, p-value = 0.001) or pfdhps K540E/N (ϕ = 0.399, p-value ≤ 0.001) mutations and pfgch1 gene amplification. Conclusions Despite withdrawal of SP as anti-malarial treatment for 17 years, the border regions of Thailand continue to display high prevalence of antifolate and anti-sulfonamide resistance markers in falciparum malaria. Significant association between pfgch1 amplification and pfdhfr (I164L) or pfdhps (K540E) resistance markers were observed, suggesting a compensatory mutation.


Background
Both falciparum and vivax malaria remains an important public health problem in border regions of Thailand. Resistance to anti-malarials presents a major hurdle for eradication of the disease [1]. Drug resistance in both Plasmodium falciparum and Plasmodium vivax has been reported as early as the 1950s [2]. Sulfadoxine-pyrimethamine (SP), a folate pathway inhibitor was deployed in Thailand for the treatment of uncomplicated falciparum malaria from 1973 until 1991 [3]. By 1991, substantial loss of the SP drug efficacy prompted a change in firstline treatment in Thailand [3]. Molecular investigations revealed that mutations in P. falciparum dihydrofolate reductase (pfdhfr) and P. falciparum dihydropteroate synthase (pfdhps) were associated with SP treatment failures and could be used as molecular markers for SP resistance [4,5]. The mutations in pfdhfr and pfdhps often occurred in a step-wise progressive manner resulting in increased levels of drug resistance [4,6]. Resistance to antifolates has also been linked to gene amplification of P. falciparum guanosine triphosphate cyclohydrolase 1 (pfgch1)-an enzyme responsible for coding a crucial enzyme in the folate pathway [7]. Parasites with pfgch1 amplification were reportedly less susceptible to antifolates as elevated expression of enzymes assisted antifolate resistance by competing with the drugs [7], and compensating the loss of fitness caused by mutations in pfdhfr and pfdhps by increasing the flux of metabolic products in the folate pathway [8]. An earlier study from Thailand reported a high proportion of parasites carrying multiple copies of pfgch1 and suggested an association between pfgch1 copy number variation (CNV) and the pfdhfr (I164L) mutations [8]. Several studies conducted between 1995 and 2008 have identified varying levels of triple or quadruple mutations in pfdhfr and pfdhps [5,8,9]. A more recent survey conducted in Ubonratchathani province close the Thailand-Cambodia borders, which had a lot of reports in many anti-malarial drug resistances [2,3], showed high levels of pfdhfr (N51I, C59R, and S108N, ≥ 76%) and pfdhps (A437G, K540E, A581G or A437G, K540N, A581G or S436A, A437G, K540E, ≥ 90%) triple mutations [10]. These border areas are malaria endemic regions. Each site is geographically distant from other and often experiences high migration of diverse human population. However, data on the current status of antifolate and anti-sulfonamide resistance markers in P. falciparum in other major border regions of Thailand is scarce. Presumably, the persistence of highly mutations on SP-resistant markers related to the using of other drugs that may also induced pressure on pfdhfr and pfdhps of malaria parasite. The trimethoprim-sulfamethoxazole, which is used to treat acute respiratory infections, presented cross-resistance with pyrimethamine and sulfadoxine [11,12]. Reemergence of chloroquinesensitive P. falciparum in Malawi after a decade-long cessation of drug use shows that for some anti-malarials restoration of drug sensitivity is possible after removal of the drug pressure [13]. However, several factors including drug target, nature of genes and host/parasite genetic background may differently affect the persistence of SP resistance after removal of SP use.
The present study is a retrospective molecular surveillance of three antifolate and anti-sulfonamide resistance markers in samples obtained from the malaria endemic border provinces of Thailand between 2008 and 2016. It aimed at describing the current status of resistance markers after long-term cessation of SP as an anti-malarial treatment in Thailand.

Study sites and samples collection
Between 2008 and 2016, dried blood spots samples were collected using Whatman 3MM CHR filter paper from P. falciparum malaria patients in nine provinces bordering Thailand (Maehongson N = 24; Tak N = 16; Kanchanaburi N = 70; Ranong N = 34; Suratthani N = 2; Ubonratchathani N = 9; Sisaket N = 19; Trat N = 12; and Yala N = 47 (Additional file 1: Table S3). All samples were acquired during anti-malarial drug therapeutic efficacy monitoring projects that were approved by the Research Ethics Committee of the Department of Disease Control, Ministry of Public Health, and Faculty of Tropical Medicine, Mahidol University, Thailand. DNA was extracted using QIAamp DNA Mini Kit (QIAGEN, Germany) and stored at 4 °C before analysis.

Plasmodium falciparum gch1 copy number estimation
To assess pfgch1 amplification, quantitative real-time PCR (qPCR) of the pfgch1 gene was compared with a single copy reference gene (seryl tRNA ligase) (Accession No: A0A0L1I323). The specific probes and primers used have been described previously [8] (Additional file 1: Table S1). A laboratory pfgch1 single copy strain and pfgch1 multiple copies strain of P. falciparum were used as reference standard. The copy number of pfgch1 gene was analysed from the threshold cycle (C t ) and ΔΔC t values calculated as follows: The copy number of pfgch1 were subsequently calculated using 2 −ΔΔCt method [18]. The pfgch1 amplification cut off value was set at 1.5 copies number.

Statistical analysis
The pfdhfr/pfdhps mutations and pfgch1 copy numbers were described as the proportion of each haplotype present at each border region. The changing trend of antifolate resistance markers after drug withdrawal was analysed by comparing the prevalence of markers over time by year. Chi-square (χ 2 ) and the non-parametric phi correlation coefficient statistics (φ) was used to evaluate the difference in frequency and associations respectively among the group of each antifolate resistance marker. A p-value ≤ 0.05 was considered statistically significant.

Mutations in pfdhfr and pfdhps genes in isolates from the Thailand-Cambodia border
Plasmodium falciparum isolates (N = 40) from malaria clinics in three provinces along Thailand-Cambodia border were collected between 2008 and 2016 (Table 1). Major haplotype was the septuple mutants that comprised of mutations at N51I, C59R, S108N and I164L in pfdhfr gene and A437G, K450N and A581G in pfdhps gene. The P. falciparum samples with septuple mutation pfdhfr-pfdhps were detected 50% (6/12)
Comparing frequencies of mutations from three areas, prevalence of septuple (ACIRNL-A/SGE/NGA) and sextuple mutations (ACIRNL-A/SGEAA, ACNRNL-A/ SGEGA and ACIRNI-A/SGE/NGA) were found in the falciparum samples from Thailand-Myanmar and Thailand-Cambodia border as the major group. Whereas, the falciparum samples of Thailand-Malaysia border presented only quintuple and quadruple mutations haplotypes. The octuple mutations which had the mutation at A/S436F or A613S/T were detected only in Thailand-Myanmar border. However, both of octuple haplotype observed below 8% in 2008-2013 were absented in 2014-2016 (Fig. 3).

Associations between pfgch1 copy number and antifolate resistance markers in pfdhfr and pfdhps
The association between individual six mutation points in pfdhfr and gene amplification of pfgch1 were assessed in all 233 samples. Only the I164L pfdhfr point mutation was significantly associated with pfgch1 amplification (φ = 0.213, p-value= 0.001). Plasmodium falciparum carrying both the pfdhfr I164 allele and pfgch1 multiple copies were observed in only 21% (15/73) of samples, while P. falciparum isolates that carried both pfdhfr 164L and pfgch1 amplification were observed in 42% (67/160) of samples (Fig. 3).
Similarly, evaluation of the association between five mutation points in pfdhps and pfgch1 gene amplification showed that the pfdhps K540E/N allele was significantly correlated to pfgch1 amplification (φ = 0.399, p-value ≤ 0.001). Plasmodium falciparum carrying the pfdhps 540E allele and pfgch1 multiple copies was observed in 50% (75/151) of samples.

Discussion
This longitudinal study of antifolate resistance markers in Thailand P. falciparum infections showed persisting high prevalence of antifolate resistance haplotypes between 2008 and 2016, despite long-term cessation of SP treatment regimen. Studies from early 1990s to 2014, revealed the pfdhfr point mutations transitioning from double mutations (S108N with N51I/ C59R/164L, ≥ 72%) to triple (N51I, C59R, S108N or C59R, S108N, I164L) and eventually quadruple (N51I, C59R, S108N/T, I164L) mutations in provinces of Thailand bordering Myanmar and Cambodia [5,10,[19][20][21]. Present findings are in agreement with previously observed trends showing that most P. falciparum parasites up to 75% carried triple or quadruple mutations in pfdhfr. However, limited data from 1994 to 1995 on the Thailand-Malaysia border indicated a high prevalence of 88% pfdhfr triple mutations, whereas no quadruple mutations were reported [22].
The pfdhps mutations based on four codons (S436F/A, A437G, A581G, and A613S/T) were first reported in mid 1990s along the Thailand-Myanmar border [5].  [9,19,23]. Taken together, previous reports [9,10] and present findings indicate emerging trend in shift from pfdhps double to triple mutations as major mutant group along the Thailand-Myanmar and Thailand-Cambodia borders.
Resistance level to antifolate drug in falciparum malaria was increased and conferred by mutations in dhfr and dhps gene. The combination of pfdhfr N51I, C59R and S108N and pfdhps A437G and K540E/N (quintuple mutant) strongly predicts clinical outcome [24]. The mutation in pfdhfr I164L and pfdhps A581G and A613S/T developed later and were associated with increase in antifolate resistance [6].
Effect of mutations in pfdhfr on the inhibition constant (K i ) for pyrimethamine has been well described [25][26][27][28]. This suggests that quadruple mutations including the 164 position were critically important for the development of resistance to pyrimethamine. In addition, it was shown that K i for sulfadoxine differed significantly between wild type and triple mutant parasites carrying K540E mutation [4,29]. Strong association between elevated copy number of the pfgch1 gene and the pfdhfr-I164L mutation has been shown previously [8]. In this study, an association between pfgch1 gene multiple copy number and the pfdhfr-I164L mutation was also observed. However, in Ghana with low level of pfgch1 gene amplification (6%; 12/192), no dhfr-I164L was detected in samples with amplified pfgch1 [30]. The positive correlation between pfdhps (540E) and pfgch1 multiple copies observed in this study was not observed before. These correlations in parasites with high-level antifolate and anti-sulfonamide resistance may indicate functional linkage and fitness epistasis between genes on different chromosomes. The observed positive correlations between pfgch1 amplification and pfdhfr (I164L)/pfdhps (K540E/N) mutations may represent adaptation of the parasite dissociated from the antimalarial drug pressure or arguably the pressure from non-anti-malarial antifolate drugs.
Present study provides indication of co-increase in quintuple mutants and pfgch1 multiple copies in five isolates during 2014-2016. However, relatively smaller sample size (in total 37 isolates) from Thailand-Malaysia border might also have influenced the observed results. Persistence of high prevalence of antifolate resistance haplotypes in Thailand may be explained by several factors. Persistence of antifolate resistance markers might result from continued drug pressure from non-malarial antifolate drugs such as trimethoprim and sulfamethoxazole [11]. Trimethoprim-sulfamethoxazole (a.k.a Bactrim ™ ) is still used in Thailand [31] as a part of standard package to care people with HIV/AIDS [32], urinary tract infection [33], melioidosis [34] and pneumonia [35]. The antifolate gene mutations might have been sustained because of continued presence of this antifolate drug pressure [11].
Human migration contributed to the spread of chloroquine resistant malaria from Southeast Asia to Africa during the 1970s, and SP resistance during the 1980s-1990s [36][37][38]. Studies conducted during 1998-2001 [39] and 2007-2009 [40], identified a high proportion of pfdhfr/pfdhps resistance haplotypes in parts of Myanmar sharing border regions with Thailand. This period coincided with the increased prevalence of pfdhfr/pfdhps mutants in Thailand [9,19,20]. The main haplotypes then were quadruple mutants for pfdhfr and triple mutants for pfdhps, respectively (Additional file 1: Table S2). The Thailand-Cambodia border showed a slight difference in the proportion of pfdhfr resistance haplotypes containing quadruple mutations, which was higher than the proportion of triple mutations observed between 2003 and 2007 [21], whereas the prevalence of pfdhps mutations was similar to Cambodia [20] (Additional file 1: Table S2). Present findings are in agreement with the low reported pfdhfr-pfdhps mutations on the Thailand-Malaysia border compared to other regions [9,22]. It is likely that the parasite population in this border area, which is a conflict zone, is separate from the other border areas, with limited human migration carrying parasites between these areas.

Conclusions
Decades after cessation of SP as anti-malarial treatment in Thailand, there is a persistent high proportion of P. falciparum carrying both sulfadoxine and pyrimethamine resistance markers, with the exception of the Thailand-Malaysia border. Amplification of pfgch1 correlated with pfdhfr (I164L) and pfdhps (K540E) mutations, suggesting pfgch1 amplification might be compensatory to mutations in the pfdhfr and pfdhps genes.
Additional file 1: Table S1. Primer and Probe sequences for detection of SP resistance markers in P. falciparum. Table S2. The prevalence reports of pfdhfr and pfdhps mutations close the border of Thailand and neighboring countries since 1990 to 2016. Table S3. The Pfdhfr/Pfdhps haplotypes and gch1 copy number variation of P. falciparum in 9 provinces along the border of Thailand-neighboring countries between 2008 and 2016.