Origins and spread of novel genetic variants of sulfadoxine–pyrimethamine resistance in Plasmodium falciparum isolates in Indonesia

Background While malaria incidence in Indonesia has decreased threefold in the last decade, more than 200,000 cases were reported in 2016. Different endemicity of Plasmodium falciparum malaria among several islands in Indonesia has been recognized and two unique mutations of P. falciparum dihydropteroate synthase (pfdhps) affecting sulfadoxine–pyrimethamine (SP) resistance were detected from the research of SP efficiency and genotype analysis in South Kalimantan. In this study, geographical distribution and origin of these pfdhps K540T and I588F mutations were analysed. Methods Malaria parasites DNA from several endemic areas in Indonesia; Sumatera, Java, Kalimantan, Lombok, Sumbawa, Timor, Sulawesi, and Papua islands; in two periods, 2004–2006 and 2009–2012 were subjected for pfdhfr and pfdhps sequence analysis. Results Different genotype polymorphisms of pfdhfr and pfdhps were observed in the parasites from various regions in Indonesia and relatively more divergent genotypes were determined from Kalimantan isolates in both 2004–2006 and 2009–2012. The parasites containing K540T mutation were identified in 2004–2006 isolates from East Kalimantan, East Java and Sumbawa as an SGTGA haplotype. The other I588F mutation was also determined in 2004–2006 parasites, isolated from Lombok and Sumbawa islands as an SGEAA(588F) haplotype. The parasites with pfdhfr/pfdhps quintuple or sextuple mutation, a genotype marker of SP resistance, were determined mostly in Kalimantan in both 2004–2006 and 2009–2012. Conclusion Analysis of the prevalence and pfdhfr/pfdhps combined genotypes of K540T or I588F mutations suggested that K540T might be origin in Kalimantan Island and I588F in Sumbawa Island and then these were spread to other areas along with people movement. This research indicates regular monitoring of drug efficacy and parasite genotype analysis is important to keep efficiency and prevent the spread of resistance. It is also essential for the latest anti-malarial drug artemisinin-based combination therapy.


Background
Malaria incidence in Indonesia has decreased threefold in the last decade [1], and there is no evidence for the presence of the parasites resistant to artemisinin-based combination therapy (ACT) [2]. However, about 25% of Indonesia's population (total population: 261 million) is at risk of malaria, and more than 200,000 positive cases were reported in 2016 [3]. Detection of drug resistant malaria parasites and prevention of these spreading are critical to keep efficacy of the malaria treatment and to obtain elimination of malaria. In this research, origin and distribution of unique mutations of Plasmodium falciparum sulfadoxine-pyrimethamine (SP) resistance in Indonesia were analysed.
SP has been widely used as an anti-malarial drug for treatment of uncomplicated malaria, and for intermittent preventive treatment in vulnerable populations, pregnant women (IPTp) and infants (IPTi) in high malaria transmission areas in Africa [4,5]. However, emergence and spread of SP resistant P. falciparum has been reported worldwide [6][7][8][9]. In Indonesia, SP was recommended as a second line anti-malarial drug after chloroquine resistance had been determined in 1973, and P. falciparum resistance to SP was reported for the first time in Jayapura (Papua Province) in 1979 [10,11]. Chloroquine and SP had been used in Indonesia until 2008, when the malaria treatment policy was changed. ACT, using a combination of an artemisinin derivative with another anti-malarial, such as piperaquine, lumefantrine or amodiaquine, is provided as the first line anti-malarial drug for treatment of uncomplicated malaria, and SP is not administered for malaria treatment. However, people, especially in local areas, use SP when they are suffering from malaria, or for chemoprophylaxis [12], because of some effectiveness, low cost, simple administration as a single oral dose, and fewer side effects. It is important to obtain information about SP resistance in malaria endemic areas in Indonesia.
The mutations in P. falciparum dihydrofolate reductase (pfdhfr) and dihydropteroate synthase (pfdhps) are responsible for pyrimethamine and sulfadoxine resistance, respectively [13][14][15][16]. Stepwise accumulation of point mutations in pfdhfr and pfdhps genes is associated with higher level of resistance to SP in vitro and in vivo [17,18]. The amino acid substitution at position 108 serine to asparagine or threonine (S108N/T) in pfdhfr is essential for subsequent A16V, N51I, C59R and I164L mutations (underlined bold type indicates the mutant allele), leading to high-level of resistance to cycloguanil or pyrimethamine [19]. Similarly, a single mutation in the pfdhps converting alanine to glycine at amino acid position 437 (A437G), which is normally the first mutation under the sulfadoxine drug pressure, conferred on the parasite a fivefold higher level of drug resistance [16]. Additional mutations K540E and A581G, then S436A/F and A613S/T are associated with increasing resistance to sulfadoxine. A combination of pfdhfr triple (N51I, C59R and S108 N) and pfdhps double (A437G and K540E) mutations collectively form the quintuple mutations, which is strongly associated with in vitro and in vivo SP resistance [17,[20][21][22][23][24].
There are several reports of mutation analysis of pfdhfr and pfdhps genes from different malaria endemic areas in Indonesia. Different ratio of mutation level was reported from sample analysis of several island and district parasites. From the West Papua sample analysis obtained in 1996 to 1999 by Nagesha et al. [25], C59R and S108N mutations in pfdhfr and A437G in pfdhps were commonly determined in SP resistant parasites. In addition, they reported an additional K540E mutation in pfdhps was observed in more resistant level parasites. Extensive analysis of the parasite genotypes from eight malaria endemic areas were reported by Syafruddin et al. in 2005, representing a broad region of the western and eastern Indonesian archipelago [26]. Polymorphisms in pfdhfr gene at S108N/T, A16V and C59R were frequently identified, in which A16V were observed in association with S108T and these were differently distributed, more common among samples from eastern regions. The polymorphism in pfdhps was less frequent in this report; about 15% of A437G and less than 5% of K540E were detected. Among the Sumba island samples of 2007, less frequent mutations were reported by Asih et al. in 2009 [27]; about 25% parasites presented S108N and C59R mutations in pfdhfr and only few % of the isolates presented A437G mutant allele of pfdhps. More prevalence of mutations in pfdhfr and pfdhps was observed from a study of SP efficacy and genotype analysis in South Kalimantan in 2009-2010 [28]. More than 90% of the isolates exhibited S108N and C59R mutations in pfdhfr gene, in addition I164L substitution was detected in 30% of the parasites. The alterations in pfdhps were detected at the amino acid positions of A437, K540, A581 and I588 to glycine (97%), glutamine or threonine (36%, 36%), glycine (45%) and phenylalanine (23%), respectively. This result of pfdhfr and pfdhps genotypes is quite unique compared with the previous Indonesia parasites; more prevalence of mutation ratio and detection of novel mutations of pfdhps K540T and I588F.
In this study, pfdhfr and pfdhps sequences of P. falciparum from different malaria endemic regions, mostly eastern part of Indonesia were analysed to investigate polymorphisms of pfdhfr and pfdhps genotypes and predict susceptibility for sulfadoxine-pyrimethamine in malaria parasites in Indonesia, and to obtain information about distribution of the previously identified pfdhps K540T and I588F novel mutations.

Study sites and malaria patients
Malaria parasites collected from the following several endemic areas in Indonesia were analysed in this study: Sumatera (Indragiri Hilir, Merangin), Kalimantan (Paser, Seruyan, Banjar), Java (Pacitan), Lombok (West Lombok), Sumbawa (Sumbawa), Timor (Timor Tengah Selatan), Sulawesi (Gorontalo), and Papua (Jayapura) islands in 2004-2006 and in 2009-2012. Malaria patients were recruited at primary health centers, district hospitals or local field areas and written informed consent were obtained from each participant, or from caretakers if participations were under 12 years old, after explanation of the purpose of the study in local language. Finger prick blood samples were collected on a glass slide for microscopical observation and on a filter paper (Advantec, Toyo Roshi Kaisha, Ltd., Japan) for parasite DNA analysis. Thick smear blood films were stained with Giemsa (Merck, Germany) and examined microscopically for the presence of malaria parasites. Dried blood spot on a filter paper was kept in a small plastic clips prior to parasite DNA extraction. In some field studies, BinaxNOW Malaria diagnosis kit (Binax Inc, Portland, ME, USA) were used to identify P. falciparum malaria patients. Patients with positive malaria diagnosis results were treated with an anti-malarial drug according to national policy. The study protocol was reviewed and approved by the Ethical Committee, Faculty of Medicine, Universitas Airlangga, Surabaya, Indonesia and Institute of Tropical Medicine, Nagasaki University, Nagasaki, Japan.

Parasite DNA analysis
Parasite DNA was extracted from the dried blood spots on filter paper by using QIAamp DNA blood mini kit (Qiagen, Netherlands) and kept at − 30 °C. Plasmodium falciparum samples were selected by nested PCR methodology using species specific primer sets of 18S rRNA genes described in Snounou et al. [29] and Kimura et al. [30]. Pfdhfr and pfdhps genotypes were determined by sequencing as previously described by Isozumi et al. [31]. using amplified PCR product as a template directly for sequence analysis. Alleles corresponding to amino acid positions at 16, 51, 59, 108 and 164 of the pfdhfr gene and at 436, 437, 540, 581, 588 and 613 of the pfdhps gene were read more carefully, and at least two independent PCR products were prepared for sequence analysis in the case of rare mutations. Previously reported results from South Kalimantan Province [28] were included in this analysis.

Statistical analysis
Data were entered in Microsoft Excel and exported to SPSS version 17.0 for analysis. Chi square and Fisher's exact tests were used, where applicable, to assess the relationship of mutations between two periods of studies. Allele proportions were calculated as the number carrying a certain allele divided by the number of samples with positive PCR outcome.

Sample characteristics
A total of 622 P. falciparum samples from symptomatic to asymptomatic malaria patients during the two periods, 2004-2006 and 2009-2012, were analysed in this study. The research areas were different levels of malaria endemicity, and presented seasonal differences in number of malaria patients. Same districts in Lombok and Papua islands, the latter is reported as high malaria endemicity, were included for sample analysis in both periods. Of the 384 samples from 622 patients, P. falciparum pfdhfr and pfdhps genes were successfully amplified and analysed both genotypes (Tables 1, 2 Polymorphisms were identified at amino acid positions of C59, S108 and I164 in pfdhfr to arginine, aspargine and leucine, respectively, but no other substitutions including at A16 and N51 were observed. In pfdhps gene, substitutions were observed at amino acid positions of A437, K540, A581 and I588 to glycine, glutamic acid or threonine, glycine and phenylalanine, respec- With the pfdhfr mutations at C59R, S108N and I164L, in total, the wild type pfdhfr allele ANCSI (at positions 16, 51, 59, 108 and 164, respectively) and three mutant alleles, ANCNI, ANRNI and ANRNL were detected in this study (Table 1).
Relatively high ratio of mutation in pfdhps gene was determined at A437 to glycine ( Fig. 1), which is regarded as the first and most essential substitution for sulfadoxine resistance [16]. Two types of mutations at amino acid position K540 were detected as AAA (lysine codon) to GAA (glutamic acid) or ACA (threonine). One of this K540T and another unique mutation I588F are recently reported alleles in Indonesia, recognized through previous South Kalimantan sample analysis [28]. The K540T was also reported in Sabah, Malaysia [32], which is located in the same island. Altogether, the wild type pfdhps haplotype SAKAA (at positions 436, 437, 540, 581 and 613, respectively) and six different mutant alleles SGKAA, SGKGA, SGEAA, SGEAA(588F), SGTG A and SGEGA were identified in this study ( Table 2) Complicated genotype variations were observed in pfdhps allele haplotypes. Similar variant haplotypes were detected from Paser (East Kalimantan) and Pacitan (Easr Table 1 Fig. 2b). This mutant was not detected in Lombok in 2009-2012.
Another unique mutation of K540T was identified as a major pfdhps SGTG A allele in Paser (East Kalimantan) and Pacitan (East Java), and in Sumbawa as a minority in 2004-2006 (Fig. 2b). This was then detected only in Kalimantan Island in 2009-2012, as a major allele of Banjar (South Kalimantan) parasites, and one case of Seruyan (Middle Kalimantan) isolates (Fig. 2b).  (Table 3). Relatively more polymorphisms in the pfdhfr/pfdhps combined genotype were observed in Paser (East Kalimantan), Pacitan (East Java), Sumbawa

Table 3 Number of cases with each pfdhfr*/pfdhps** combined genotype* from different districts in Indonesia in two study periods
Year Accumulation of mutations in pfdhfr and pfdhps genes enhances parasite resistance level against sulfadoxinepyrimethamine. The parasites containing more than five mutations in combined pfdhfr/pfdhps genotype have been shown not to respond adequately to SP treatment [8,[22][23][24]. Prevalence of the quadruple ANRNI/SGEAA(588F) genotype, and quintuple or sextuple mutant genotypes in each research area is presented in Fig. 3 (Table 4).

Discussion
In but pfdhfr A16V and S108T mutations for cycloguanil resistance were not detected (Fig. 1). In pfdhfr gene, wild type ANCSI and three mutant ANCNI, ANRNI and ANRNL alleles were observed (Table 1). This simple accumulation pattern of mutations in the pfdhfr genotype supports the stepwise selection hypothesis of resistant gene evolution [19]. Wild type SAKAA and six different mutant alleles were determined in pfdhps gene ( Table 2). Comparison of wild type and each mutant alleles suggests stepwise accumulation of mutations in the pfdhps genotypes.
The unique K540T and I588F mutations of pfdhps, both of which were detected previously from 2009 to 2010 South Kalimantan sample analysis [28] and the former was also reported by Lau et al. [32] from 2010 Sabah, Malaysia parasites, were identified in 2004-2006 parasites. The evidence presenting here that K540T was detected in 2004-2006 as an SGTG A haplotype from several parasites in East Kalimantan, East Java and Sumbawa Island suggests a possibility of Kalimantan Island origin for this mutation. In East Kalimantan, the SGTG A was found from many patients as a combination genotype of either ANRNI/SGTG A or ANRNL/ SGTG A, additionally as a mixed infection of both genotypes (Table 3). Meanwhile, the SGTG A existed as the combination genotype of ANRNI/SGTG A only in East Java and Sumbawa parasites.
Another novel mutation I588F was also detected in 2004-2006 samples from Lombok and Sumbawa islands; the parasites of SGEAA both with and without I588F mutation were detected in Sumbawa, whereas SGEAA(588F) and wild type SAKAA were observed in Lombok. It suggests a possibility that the initial mutation of I588F had occurred in SGEAA type parasites in Sumbawa Island, then introduced into its neighbor Lombok Island. Further analysis and comparison of microsatellite loci around the pfdhps haplotypes will provide additional information for understanding the origin and spreading of these K540T and I588F mutant alleles.
In Riau and Jambi Provinces (Sumatera Island), only small number of malaria patients had been detected and these were mostly infected with P. vivax and only a few P.    Fig. 2b) and many cases of mixed genotype infections (Table 3). This suggests high P. falciparum infection rate under strong pressure of SP. The mixed genotype infection was not common in 2009-2012 samples from Middle and South Kalimantan, and the parasites from these Provinces presented characteristic genotype polymorphisms. Most of the parasites from Seruyan (Middle Kalimantan) acquired pfdhfr I164L mutation (ANRNL haplotype, Fig. 2a), and several pfdhps genotypes involving unique pfdhps K540T or I588F mutations were detected in Banjar (South Kalimantan) parasites (Fig. 2b).
High heterogeneity of malaria epidemiology and divergent genetic polymorphisms across islands, districts, and even close neighbour sub-districts in one island are not uncommon in Indonesia [26]. Divergent polymorphisms in pfdhfr/pfdhps genotype of the parasite populations from different Kalimantan provinces were observed in this study from analysis of the parasite samples before introduction of ACT at each research site in these Provinces. AA was implemented in 2006 in East Kalimantan and then DHP treatment was started in 2009. DHP has been first-line treatment therapy against malaria in Middle Kalimantan and South Kalimantan since 2010. However in fact, ACT was introduced gradually form one site to the other, and the parasite samples from Kalimantan analysed in this research were collected before ACT was applied in the research areas. Current situation and comparison of genetic divergence among Kalimantan provinces are important subject to provide information how application of ACT influences malaria parasite populations on genotypes and polymorphisms. Different prevalence of quintuple or sextuple mutant parasites in pfdhfr/pfdhps combined genotypes were observed (Fig. 3). While the parasites from Kalimantan and Pacitan (East Java) belonged to SP resistant quintuple or sextuple mutant genotype, parasites from the other areas in Indonesia presented four or less mutations in combined genotype that tend to adequate response for SP treatment. In addition, efficacy of SP treatment for P. vivax malaria infection in Indonesia was reported by Asih et al. [34] recently. These suggest SP could be considered for prevention or treatment of malaria as a single prescription or in combination with artemisinin in Indonesia except in Kalimantan Island. In such a case, regular monitoring of the efficacy and regular genotype analysis are essential to prevent spread of resistance.
Mobility of people among thousands of islands is one of the important factors for malaria control in Indonesia. Some of outbreaks, resurgences of malaria had occurred under the unique circumstances of people migration. Marwoto et al. reported immigrant workers who worked as transmigrants or seasonal workers in malaria endemic areas outside Java Island returned to their home villages brought imported malaria cases [35,36]. Many migrants and temporal workers have moved from Pacitan (East Java) to several islands historically. Similar pfdhfr/pfdhps genotype polymorphisms in 2004-2006 parasites from Pacitan (East Java) and 2009-2012 Saruyan (Middle Kalimantan) in Table 3 suggests the malaria parasites in these districts could have been transferred along with human migrations.
The Indonesia National Malaria Control Programme desires to eliminate malaria in the whole country by 2030 [37]. Most districts in Java and Bali islands, also several districts in other islands fall under the World Health Organization criteria of elimination and malaria cases have gradually decreased over the last several years in Indonesia [38].