Introduction

The emergence of artemisinin resistance is of great concern since the current first-line treatment for Plasmodium falciparum malaria in endemic countries relies on artemisinin-based combination therapies (ACT) [1]. Partial artemisinin resistance has been defined as a slower parasite clearance rate following treatment with an artesunate monotherapy, or after treatment with an (ACT) [1] and is increasingly common in South East Asia. Nevertheless this definition is affected by multiple factors such as patient immunity, blood drug concentration or partner drug activity [1]. Clinical response to ACT treatments (treatment failures) reflects the efficacy of both drugs in the combination, being the proportion of patients remaining parasitemic on day 3 the indicator of choice, currently used in routine monitoring for suspected artemisinin resistance [1]. Treatment failures observed in South East Asia correlate either with lower efficacy of the partner drug and with slower parasite clearance rate due to artemisinin partial resistance [2]. Historical evidence shows that the emergence of drug-resistant (chloroquine and sulphadoxine-pyrimethamine) P. falciparum strains first occurred in Southeast Asia, spread to East Africa and then spread outwards to the rest of the African continent [3].

During the last decade, African countries have gradually changed first-line treatment for uncomplicated falciparum malaria to ACT due to the development of resistance to the successively introduced anti-malarial drugs. In 2002, Mozambique (East Africa) replaced chloroquine by sulphadoxine-pyrimethamine in combination with amodiaquine (AQ-SP); in 2004 artesunate substituted AQ (AS-SP) and since 2009, this combination changed to artemether-lumefantrine [4]. In Angola (West Africa), the artesunate-amodiaquine combination was officially adopted in 2004, followed by artemether-lumefantrine in 2006 [5] and ACTs have been used nationwide since 2007.

In 2014, Ariey et al. [6] associated the artemisinin resistant phenotype to mutations on the kelch propeller protein K13 encoded in PF3D7_1343700 gene. In Cambodia, where these polymorphisms were first described, the presence of K13 mutants (mainly Y493H, R539T, I543T and C580Y) are correlated to in vitro parasite survival rates and in vivo parasite clearance rates. In August 2014, the World Health Organization included K13 mutations in the new working definition for partial artemisinin resistance stating that the presence of ≥5% of patients carrying K13 resistance-associated single nucleotide polymorphisms (SNPs) defines suspected artemisinin resistance; when K13 mutants are associated with either persistent parasitemia on day 3 or longer parasite clearance, resistance is confirmed [1].

K13-propeller SNPs have been searched in multiple locations aiming to describe the presence of resistant parasites before clinical resistance spreads, following the recommendations of WHO. To date, more than 60 mutations in K13 gene have been described in Africa and South East Asia [2,6,7–10], however, no resistance-associated mutations have been found outside the Greater Mekong Area [1].

In this study we report the presence of polymorphisms in P. falciparum K13-propeller in Angola and Mozambique, two portuguese-speaking countries endemic for malaria, before and after the introduction of ACT as first-line treatment.

Methods

Results

P. falciparum K13-propeller polymorphisms observed

A total of 3 different mutations (R471R, V494I, R575R) were found in 5 samples (2.5%). All mutations were found in samples collected after the introduction of ACTs as the official first-line treatment in each country. Fig. 1 shows geographical origin of the samples and time period, and Table 1 shows K13-propeller polymorphisms identified.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. P. falciparum K13-propeller polymorphisms prevalence and distribution in Angola and Mozambique.

Panel 1) Proportion within each time period of wild type (wt) and mutant (mut) samples; Panel 2) Geographical location of sample collection (dark grey) and identified polymorphisms.

https://doi.org/10.1371/journal.pone.0119215.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. P. falciparum K13-propeller amino acid and nucleotide substitutions observed in Angola and Mozambique.

https://doi.org/10.1371/journal.pone.0119215.t001

The newly described polymorphism V494I was found in two samples from Maputo province (Mozambique). Two synonymous polymorphisms (R471R in two samples and R575R in one sample) were found in Angola. R575R, also newly described, was found in Malanje, an Angolan province that borders the Democratic Republic of Congo. R471R (found in Luanda and Malanje) has already been observed in DR Congo (Kinshasa) [9] and Gabon (Libreville) [10]. All polymorphisms were uncommon (R471R 4%, V494I 4%, R575R 2%), and each was present in different isolates. The mutations associated to artemisinin-resistance (Y493H, R539T, I543T and C580Y) described in Southeast Asia [2,6] were not detected.

Discussion

The shift in drug policy in Africa towards ACTs was recent, it has been accompanied by both the selection of polymorphisms associated with decreased lumefantrine sensitivity [13,14] and by the decrease of the in vitro sensitivity to this drug [15]. The reduced sensitivity to ACT partner drugs may increase the selection of artemisinins resistant parasites. Thus, the use of the new molecular marker K13-propeller is fundamental for surveillance in malaria control programs, in order to prolong the life span of the ACTs in Africa.

Our study on P. falciparum K13-propeller polymorphisms showed the presence of three mutations, two of them not previously described. One of the mutations is a non-synonymous SNP (V494I) that lays one codon next to the known K13 resistance-associated mutation (Y493H). Although functional analysis for V494I did not predict deleterious effect on protein function (PROVEAN score = -0,774) as it does for Y493H (score = -4,166), this does not necessarily mean that the protein is not affected. For instance, R539T was considered neutral in a similar analysis (score = -2,489), although it is associated to ACT drug response in Southeast Asia [6]. Plus, V494I was identified in two samples collected in Mozambique after the introduction of ACTs. Mozambique is located at East Africa, a region recognized as being a major focus of drug resistant parasites entrance and development in sub-Saharan Africa, probably originated at Southeast Asia [3]. While speculative, justifies further attention. The other two polymorphisms described in our study encode synonymous substitutions, thus not affecting protein structure and function. Mutation R471R observed in samples collected after ACT introduction in Malanje and Luanda provinces (Angola) has also been reported in two other western African countries, DR Congo [9] and Gabon [10], however it is difficult to attribute this to the spread of a P. falciparum clone or to multiple emergence of this SNP. Nevertheless, parasites in endemic regions of sub-Saharan Africa are less likely to be clonal because of the high transmission intensity [10]. Recently, K13 resistance-associated mutations have been demonstrated to arise both population-specific and independently in different geographic areas in Southeast Asia [2]. Multiple non-synonymous polymorphisms have been published, but there are very few shared mutant alleles between African and Southeast Asia parasites as yet [2,6,9].

The absence of K13 resistance-associated mutants from Southeast Asia in Africa [7–10] is promising, but continuous tracking of resistance emergence should be implemented enabling early detection of resistant parasites. The use of a molecular marker such as K13 mutations is nowadays a cornerstone for malaria surveillance programs, but differences between African and Asian K13-mutant parasites should be taken into account, and this study provides useful information on two not previously described K13 polymorphisms in Angola and Mozambique where malaria is a serious public health problem.

Whether or not the mutations identified during the present study are associated to drug response, can not be inferred from our results (no phenotype of drug resistance was observed in the mutant carrier samples). Nevertheless, no SNPs were present in the samples collected before the introduction of the ACTs as a first-line of treatment in each country.