Background

Malaria is a significant public health problem because of its worldwide distribution and high mortality. It was estimated that there were 219 million cases of malaria globally in 2017, mostly in 15 sub-Saharan African countries and India, representing approximately 80% of the global malaria burden [1]. Of the five malaria species, Plasmodium falciparum caused the most malaria incidence worldwide, accounting for 99.7% of estimated malaria cases in the World Health Organization (WHO) African region and 62.8% in Southeast Asia in 2017, and is the causative agent of the most severe forms of the disease [1].

Effective treatment is critical. Since the 1940s, multiple anti-malarial drugs have been developed and used to treat malaria parasites, including chloroquine (CQ), mefloquine, quinine, pyrimethamine, and sulfadoxine. However, the widespread use of these drugs promotes drug resistance, especially chloroquine resistance (CQR). Resistance to CQ occurred in Southeast Asia, South America, and the Western Pacific region in the late 1950s and rapidly spread to malaria-endemic areas, including Africa [2, 3]. Mutations in the chloroquine resistance transporter (Pfcrt) located on the P. falciparum digestive vacuole membrane were responsible for CQ treatment failure [4, 5]. Amino acid polymorphisms at PfCRT amino acid residues 72–76 were observed in CQR field isolates, whereas CVMNK haplotypes at PfCRT residues 72–76 were regarded as chloroquine sensitive (CQS) [6, 7]. Other studies revealed that Pfcrt K76T variants could affect parasite fitness, increase the rate of gametocyte production, and alter the susceptibility to artemisinin-based combination therapy (ACT) [3, 8,9,10]. These results highlight the need to monitor the molecular evolution of Pfcrt.

In view of P. falciparum multidrug resistance, the WHO recommended ACT as the first-line treatment for uncomplicated P. falciparum malaria in 2006 [11]. However, the detection of artemisinin-resistant P. falciparum in western Cambodia and the border between Cambodia and Thailand in 2008 was a drawback to malaria elimination [12, 13]. Over 200 nonsynonymous P. falciparum Kelch13 (k13) mutations have been reported to date, of which nine variants (F446I, N458Y, M476I, Y493H, R539T, I543T, P553L, R561H, and C580Y) were correlated with slow parasite clearance and reduced in vitro drug sensitivity, and over 20 k13 mutations are considered candidates or associated markers [14]. k13 mutations were detected predominantly in the Greater Mekong subregion(GMS) [15]. k13 mutations are rare in Africa, and their profile is highly heterogeneous [16]. The prevalence of nonsynonymous k13 mutations is low in approximately 50% of African countries [15]. Nevertheless, the increase in drug resistance in Africa could hamper malaria control considering its high morbidity and mortality. Therefore, monitoring mutations associated with artemisinin resistance via delayed parasite clearance globally, but especially in Africa, is critical.

Zhejiang province, located in eastern China, was considered malaria-free in 2018. No indigenous malaria infections have been reported in Zhejiang province since 2011. However, approximately 200 cases are imported every year, especially P. falciparum malaria from Africa. In this study, samples were collected from P. falciparum cases imported into Zhejiang Province, China, between 2016 and 2018, and molecular surveillance of Pfcrt and k13 was performed to determine the emergence and spread of drug resistance in the countries of origin.

Methods

Sample collection and DNA extraction

Zhejiang province located in the Yangtze River Delta region and has many migrant workers travelling from Africa and Southeast Asia each year. Since 2012, all reported cases are imported. Microscope examination and PCR targeting the DNA of the P. falciparum multicopy 18S ribosomal RNA gene were employed to confirm suspected infections.

In this study, 485 cases of P. falciparum imported into Zhejiang Province, China between January 2016 and December 2018 were investigated. Venous blood was obtained from each patient. In total, 485 whole blood samples were collected. All blood samples were stored at − 80 °C until use. Parasite genomic DNA was extracted from 200 μL of blood using the QIAamp DNA Mini kit (QIAGEN Inc., Germany) following the manufacturer’s instructions.

DNA amplification and sequencing

The k13 and Pfcrt genes were amplified by nested PCR, as previously described [17, 18]. The primers for PCR were described in previous study and they are listed in Additional file 1: Table S1 [17, 18]. Proof-reading polymerase was used in each reaction to prevent amplification errors. The proof-reading polymerase was contained in the High Fidelity PCR Master Mix (item number B639292), which was supplied by Sangon Biotech Co., Ltd. (Shanghai, China). The amplification conditions in both PCR rounds were as follows: one cycle at 95 °C for 2 min, followed by 30 cycles at 95 °C for 2 s, 60 °C for 90 s, and 72 °C for 90 s, and one extension cycle at 72 °C for 10 min. PCR products were sequenced by Sangon Biotech Co., Ltd. (Shanghai, China).

Data analysis

Mega version 7.0.26 (https://www.megasoftware.net/) was used to align amplicon sequences to 3D7 reference strain sequences retrieved from the NCBI database. A database was constructed using Microsoft Excel 2017, and statistical analysis was performed with SPSS Statistics for Windows version 21.0 (IBM Corp., Armonk, NY, USA). P-distance, the proportion of nucleotide sites at which two sequences being compared are different, was obtained by dividing the number of nucleotide differences by the total number of nucleotides compared. It was estimated by Mega version 7.0.26. Nucleotide diversity (Jukes and Cantor) (Pi (JC)), the average number of nucleotide substitutions per site between two sequences, was obtained using the Jukes and Cantor correction. Tajima’s D test was used as a neutrality test. Number of segregating sites (S), haplotype diversity (Hd), Pi (JC), average number of nucleotide differences (k) and Tajima’s D test were performed using DnaSP 6.12.03. Pearson Chi-square test was used for statistical analysis. Variables with a P value smaller than 0.05 were considered statistically significant.

Results

General information

A total of 485 P. falciparum cases imported from 37 countries from 2016 to 2018 were included in this study. Most cases involved migrant workers returning from Africa. Blood samples were collected from each patient, and 192, 169, and 124 samples were collected in 2016, 2017, and 2018, respectively (Additional file 1: Table S2). The median (range) age of the study population was 42 (9–69) years. A total of 426 out of 485 patients were men. Most cases were imported from West Africa (54.43%, 264/485), Central Africa (24.54%,119/485), and South Africa (12.16%, 59/485) (Fig. 1). Only a few cases originated in East Africa, North Africa, Philippines, Myanmar and Papua New Guinea, accounting for 6.80% (33/485), 0.82% (4/485), 0.21% (1/485), 0.21% (1/485) and 0.82% (4/485), respectively.

Fig. 1
figure 1

The number of imported P. falciparum cases from Africa and percentage of haplotypes of Pfcrt

Full size image

Pfcrt mutations

The amplified 145 bp fragment of Pfcrt encoded amino acid residues 72 to 76. Of the 485 samples, 471 were sequenced successfully, whereas 14 isolates [DR Congo (n = 2), Gabon (n = 2), Congo (n = 1), Mozambique (n = 1), Angola (n = 2), Malawi (n = 1), Republic of South Africa (n = 1), Guinea (n = 1), Nigeria (n = 3)] were not sequenced successfully because of the poor quality of DNA (Table 1). Wild-type Pfcrt alleles (CVMNK), mutant Pfcrt alleles (CVIET, SVMNT) and mixed Pfcrt alleles (CVM/I N/E/D/K K/T) were detected, and the most prevalent alleles were wild-type CVMNK (72.61%, 342/471) and mutant (22.72%, 107/471). Of the mixed haplotypes, the nucleotide sequences corresponding to residues 74–76 were ATG/T, A/GAA/T, and AA/CA, and could encode different amino acid sequences (CVM/I N/E/D/K K/T) (Table 1, Fig. 2).

Table 1 Prevalence of Pfcrt genotypes between 2016 and 2018
Full size table
Fig. 2
figure 2

Electropherograms of k13 and Pfcrt genotypes detected in imported P. falciparum cases

Full size image

Spatial distribution of Pfcrt mutations

Pfcrt mutations were more common in patients from West Africa (26.92%, 70/260), North Africa (25.00%, 1/4) and Central Africa (21.93%, 25/114). The smallest percentage of Pfcrt mutations was found in East Africa (6.06%, 2/33). The difference in the prevalence of Pfcrt mutations between West, Central, North, East, and South Africa was statistically significant (χ2 = 14.165, P < 0.05). The spatial distribution of these mutations was highly heterogeneous. The percentage of wild-type, mutant, and mixed haplotypes varied greatly between countries. Although the number of cases originating in Asia and Oceania was small, these cases had a high rate of the mutant type. One out of two patients returning from Asia presented mutant alleles CVIET and three out of four patients returning from Papua New Guinea harbored SVMNT haplotypes (Table 1, Fig. 1).

k13 propeller mutations

The nested PCR yielded an 849-bp amplification product. Among the 485 samples, 437 were successfully amplified and sequenced (Central Africa:106, East Africa: 21, North Africa: 4, South Africa: 53, West Africa: 247, Southeast Asia: 2, Oceania: 4). Twenty-six samples presented 19 different point mutations, including eight nonsynonymous mutations (P441S, D464E, K503E, R561H, A578S, R622I, V650F, N694K) (Table 2). DNA sequences with the 19 point mutations were submitted to the NCBI database under GenBank accession numbers MN586239 to MN586257. It is worth noting that an R561H mutation, one of the nine validated molecular markers associated with for artemisinin resistance via delayed parasite clearance, was detected in two patients, one each from Myanmar and Rwanda. Moreover, two samples harboring a P441S mutation contained either mutant or mixed Pfcrt haplotypes (Table 2). The A578S substitution, although common in Africa, was only found in one patient returning from Cameroon, and K503E was identified in another sample. In addition, R622I was detected in two samples, one each from Mozambique and Somalia. V650F, N694K, and D464E substitutions were present in parasite strains from Nigeria, Côte d’Ivoire, Angola, and Guinea, respectively.

Table 2 Mutant types observed in the k13 of imported P. falciparum infections
Full size table

Genetic diversity of k13

The genetic diversity of the k13 gene represented by the P-distance, S, Hd, Pi, and k values was low in the imported samples (Table 3). Similarly, only 15 haplotypes were observed in isolates imported from West Africa, with a haplotype diversity of 0.141. The patients from East Africa only had two haplotypes, but the haplotype diversity, nucleotide diversity, and average number of nucleotides were the highest compared with those from other areas. After combining all the samples, Tajima’s D significantly deviated from neutrality. However, when the samples from the different regions were analysed separately, the negative Tajima’s D on k13 was only evident in the isolates from West Africa. The Tajima’s D values for the samples from South Africa, Central Africa, and East Africa were not statistically significant, but they exhibited a similar trend to that obtained for the samples from West Africa.

Table 3 Genetic diversity of k13 gene from imported P. falciparum infections
Full size table

Discussion

As one of the most important approaches to combat malaria, chemotherapy is paramount to treat Plasmodium and interrupt malaria transmission. However, effective anti-malarial drug policies were followed by drug resistance, which temporarily interrupted malaria elimination campaigns. Drug resistance has posed a significant threat to global malaria control strategies and raised international concern, especially in areas most strongly affected by the disease [19]. Most malaria cases originated in the WHO African region, accounting for approximately 92% of all cases and 93% deaths [1]. Four countries in Africa represented almost 50% of all malaria cases worldwide: Nigeria (25%), The Democratic Republic of Congo (11%), Mozambique (5%), and Uganda (4%) [1]. Furthermore, African countries spread the infection to other countries, including France (> 2500), China (> 2000), UK (approximately 1500), and Germany (> 500) in 2017 [1, 20]. Zhejiang province, located in southeast China, had several cases of P. falciparum malaria imported from Africa in recent years [21, 22]. The current study evaluated drug resistance markers of P. falciparum infections predominantly from Africa using molecular assays to advance drug policies against malaria.

Chloroquine, an easy-to-use and affordable first-line antimalarial agent, is comprehensively used globally to treat P. falciparum and Plasmodium vivax. Nevertheless, this drug was withdrawn from most endemic countries due to the high levels of resistance, which has resulted in a two- or threefold increase in malaria deaths and hospital admission for severe malaria in various African countries [19, 23]. Over ten multi-mutations sharing the common K76T substitution have been detected and are associated with CQR in field and laboratory strains of P. falciparum [4, 24, 25]. Five major haplotypes at Pfcrt residues 72–76 (CVIET, SVMNT, SVIET, CVMNT, and CVTNT) were related to CQR, and CVIET and SVMNT were regarded as the most resistant haplotypes [7, 26,27,28]. CVIET is predominant in Africa, whereas SVMNT is more common in South America [27]. Our study confirmed that CVIET was the most common mutation type in infections imported from Africa, similar to previous studies [26, 28]. The high frequency of SVMNT in Papua New Guinea was also consistent with other studies [29]. It was demonstrated that the spatial distribution of mutant alleles was mainly related to local drug policy. The increase in the prevalence of SVMNT isolates in Tanzania from 0% (0/156) in 2003 to 3.68% (6/163) in 2004 was due to selective pressure for amodiaquine resistance [27]. Therefore, it was postulated that CVIET might be displaced by SVMNT in African regions where amodiaquine was increasingly used [27, 30]. However, our study refuted this hypothesis by showing the absence of SVMNT double mutants in Tanzania and other African countries. The disagreement in the results may be due to differences in sample size and survey sites. The present results indicate that parasites with CVIET are still a major threat in Africa.

With the cessation of CQ use between 1998 and 2008, parasite populations with wild-type CVMNK returned progressively, demonstrated by the increased frequency of this haplotype in Africa [30,31,32,33]. For instance, the frequency of a CQS genotype increased from 28.0% in 2003 to 53.7% in 2012 after disuse of CQ for 9 years in Cameroon [34]. This hypothesis is supported in our study by the consistent detection of CVMNK (72.61%) in imported malaria cases and by another study wherein most P. falciparum isolates from Africa harbored wild-type alleles [33]. A possible explanation is that CVIET mutants were not fit enough to evolve into wide-type strains without selective pressure [30]. These results help understand the dynamics of major Pfcrt haplotypes in Africa and enrich the evidence for drug policy.

K13 propeller polymorphisms in P. falciparum have been widely studied since it was first described in GMS [12]. Molecular markers for k13 have been identified and can help elucidate artemisinin delayed parasite clearance [15]. In this study, 19 alleles were identified, and eight (P441S, D464E, K503E, R561H, A578S, R622I, V650F, and N694K) were nonsynonymous. Similarly with previous literatures, the propeller domain of the k13 gene showed a limited diversity of alleles in Africa [18, 35, 36]. Of note is that one of the validated k13 resistance mutations—R561H—was detected in two patients, one each from Rwanda and Myanmar. R561H in Myanmar and western Thailand, although not the predominantly popular, was more prevalent than that in Africa [33, 37, 38]. The only infected patient from Myanmar was positive for R561H. Previous studies found the R561H allele in Congo. However, R561H might be the first time found from Rwanda [36, 37, 39]. In addition, mutation A578S identified in one isolate from Cameroon, was also found in other African countries, including Gabon, Uganda, Mali, Kenya and DR Congo [18, 39,40,41], but it was not associated with clinical or in vitro resistance to artemisinin according to previous studies [37, 42]. The study also found the uncommon k13 mutation R622I in two cases, one in Mozambique and one in Somalia, and this mutation was initially reported in northwest Ethiopia at the border with Sudan [43]. A previous study showed that, of a total of over 14,000 screened samples from 59 countries in Africa, Asia, Oceania, and South America, only one sample from Zambia was positive for R622I [37]. Interestingly, one out of three patients from Ethiopia bearing R622I mutation showed day-3 positivity in Giemsa-stained smears [43]. Further clinical studies are required to investigate the role of R622I in acquired resistance to artemisinin.

In Nigeria, where nearly 25% of the samples collected in this study originated, seven samples exhibited six mutant types (two nonsynonymous types comprising P441S and V650F, and four nonsynonymous types comprising C469C, V589V, N664N, and A676A). Similarly, no validated or candidate amino acid mutations were detected in southwestern Nigeria according to recent studies [35, 44]. However, a low prevalence of single nucleotide polymorphisms (G665C, V666V, P553P, V510V, A578S, D464N, and Q613H) were also reported, although they differed from the results obtained in the current study [35, 44]. In general, these results indicate that the profiles of the molecular markers conferring artemisinin delayed clearance were still optimistic with respect to the public health situation in terms of malaria in Nigeria.

The molecular genetic analyses conducted in this study showed that the haplotype diversity of the samples imported Africa was relatively low. Similarly, haplotype diversity of P. falciparum isolates in Congo, Ghana, Kenya and Tanzania in a previous study was 0.067, 0.123, 0.066 and 0.056, respectively [45]. Also, southwestern Nigeria reported Hd 0.080–0.157 in two states [35]. However, another research reported Hd 0.74 from African isolates [46]. The difference might result from heterogeneity of spatial distribution of samples. The current study also showed that the sequence diversity varied among the isolates from West, South, East, and Central Africa. The isolates from East Africa had the highest genetic diversity, whereas those from Central Africa had the lowest diversity. The discrepancy might have been related to the unequal sample sizes in different areas. Further investigation is required. Regarding the neutrality test, there was no consistent results. Tajima’s D value in this study was found to be negative and statistically significantly in terms of the total samples from Africa. Another research also reported similar trends from African isolates [46]. Nevertheless, negative Tajima’s D value without statistical significance was observed in Nigeria which indicated k13 gene evolved under neutral model [35].

Conclusions

In conclusion, this study determined the frequency and spatial distribution of Pfcrt and k13 mutations associated with drug resistance in P. falciparum malaria cases imported into Zhejiang province. In Africa, wild-type Pfcrt was predominant, but detection of k13 mutants was limited by the absence of genetically validated molecular markers of artemisinin delayed parasite clearance, except one case from Rwanda harbouring the R561H mutation. Both Pfcrt and k13 mutations were detected in patients from Southeast Asia and Oceania. These results may guide efforts to make more rational and targeted drug policies to eliminate resistant malaria in the study region and demonstrated that molecular markers are an effective and convenient tool to improve the surveillance of drug resistance.