Tracking origins and spread of sulfadoxine-resistant Plasmodium falciparum dhps alleles in Thailand

Abstract

The emergence and spread of drug-resistant Plasmodium falciparum have been a major impediment for the control of malaria worldwide. Earlier studies have shown that similar to chloroquine (CQ) resistance, high levels of pyrimethamine resistance in P. falciparum originated independently 4 to 5 times globally, including one origin at the Thailand-Cambodia border. In this study we describe the origins and spread of sulfadoxine-resistance-conferring dihydropteroate synthase (dhps) alleles in Thailand. The dhps mutations and flanking microsatellite loci were genotyped for P. falciparum isolates collected from 11 Thai provinces along the Burma, Cambodia, and Malaysia borders. Results indicated that resistant dhps alleles were fixed in Thailand, predominantly being the SGEGA, AGEAA, and SGNGA triple mutants and the AGKAA double mutant (mutated codons are underlined). These alleles had different geographical distributions. The SGEGA alleles were found mostly at the Burma border, while the SGNGA alleles occurred mainly at the Cambodia border and nearby provinces. Microsatellite data suggested that there were two major genetic lineages of the triple mutants in Thailand, one common for SGEGA/SGNGA alleles and another one independent for AGEAA. Importantly, the newly reported SGNGA alleles possibly originated at the Thailand-Cambodia border. All parasites in the Yala province (Malaysia border) had AGKAA alleles with almost identical flanking microsatellites haplotypes. They were also identical at putatively neutral loci on chromosomes 2 and 3, suggesting a clonal nature of the parasite population in Yala. In summary, this study suggests multiple and independent origins of resistant dhps alleles in Thailand.

FIG. 1.

Region-wise distribution of dhps alleles in Thailand. The gray-shaded areas in the map indicate the 11 Thai provinces from where samples were collected (Mae Hong Son, n = 38, including 5 samples from Chiangmai; Tak, n = 151; Kanchanaburi, n = 52; Prachuap Khiri Khan, n = 37; Ranong, n = 55, including 13 samples from Chumphon; Yala, n = 46; Trat, n = 24, including 10 samples from Chanthaburi; Si Sa Ket, n = 14). Pie diagrams illustrate the proportion of various dhps alleles. For comparison, the distribution of dhps alleles in the Pailin (n = 29) and Chumkiri (n = 49) regions of Cambodia are also shown (46). The histogram (inset) shows the frequencies of three major triple mutant dhps alleles, AGEAA, SGEGA, and SGNGA, in the Thailand-Cambodia region (dhps codons 436, 437, 540, 581, and 613 are sequentially represented, with mutated codons underlined). MH, Mae Hong Son; CM, Chiangmai; TK, Tak; KN, Kanchanaburi; PC, Prachuap Khiri Khan; RN, Ranong; CP, Chumphon; YL, Yala; TR, Trat; CB, Chanthaburi; SS, Si Sa Ket; PL, Pailin; CK, Chumkiri.

FIG. 2.

Selection valley around dhps alleles in Thailand. (A) The genetic diversity of the isolates from Yala (n = 30) were calculated separately from the remainder of the Thai isolates (n = 270). The diversity was extremely reduced at all 10 dhps microsatellite loci (mean H_e, 0.02 ± 0.01) in Yala and was not significantly different (P = 0.76) from the mean H_e at neural loci (0.05 ± 0.03). The selection valley for the rest of the Thai isolates was deep, and the mean H_e at dhps loci (0.42 ± 0.06) was very low (P = 0.0002) compared to the mean H_e at neutral loci (0.82 ± 0.04). (B) Shape of the valley around three triple mutant dhps alleles, AGEAA, SGEGA, and SGNGA. The mean H_e values around these alleles were not significantly different from each other. The error bars indicate ±1 standard deviation (SD).

FIG. 3.

The 10-locus microsatellite haplotype profiles (T1 to T98) flanking dhps in Thailand. Each box represents the occurrence of the dhps allele (codons 436, 437, 540, 581, and 613 are sequentially represented, with mutated codons in red) and its flanking microsatellite haplotype in the 11 Thai provinces. The bar graphs under the column “total” indicate the number of isolates with that particular haplotype. Identical colors (pink, green, and blue) represent proposed common lineages. Box I represents one major lineage for the AGEAA allele and its minor variants. Box II represents lineages of the AGKAA allele. The AGKAA lineages are closely related to some of the haplotypes of AGEAA, as can be seen by the sharing of blue and pink between boxes I and II. Boxes III and IV represent one major and several minor lineages of SGEGA and SGNGA, respectively. As seen here, SGEGA and SGNGA have common origins independent from that of AGEAA. The haplotypes around SGKGA alleles (box V) were also identical or nearly identical to the haplotypes around SGEGA/SGNGA. Boxes VI to IX represent low-frequency dhps alleles with their microsatellite haplotypes. For additional information on the genetic relationships among these 98 haplotypes, refer to the neighbor-joining tree shown in Fig. 4. Haplotypes common for different dhps alleles were given the same name. As an example, T1 is the major haplotype for AGEAA (box I) but was also seen for one isolate bearing SGEGA (box III). Haplotype T37 was common for the SGEGA, SGNGA, and SGKGA alleles. We used the eBURST (version 3) algorithm to construct these 10-locus microsatellite haplotypes (17), with some manual formatting. H, haplotype; ND, allele sizes not determined.

FIG. 4.

NJ tree depicting relationships between dhps microsatellite haplotypes in Thailand and Cambodia. T1 to T98 represent haplotypes for 300 Thai P. falciparum isolates, as shown in Fig. 3. Blue (AGEAA) and black (SGEGA/SGNGA) branches indicate two major independent lineages of the triple mutants. Red and green branches indicate the SGEAA and AGKAA alleles, respectively. All other minor dhps alleles are indicated. For comparison, we also included 33 dhps microsatellite haplotypes deduced from 78 Cambodian isolates (46). Thirteen Cambodian haplotypes (representing 54 isolates) were identical to the haplotypes found in Thailand (branch tips marked as gray circles) and were thus given the same name (Table 3). The remaining 20 Cambodian haplotypes (representing 24 isolates) were also very similar to some of the Thai haplotypes and are indicated as C1 to C20 in this figure.