Anti-malarial resistance in Mozambique: Absence of Plasmodium falciparum Kelch 13 (K13) propeller domain polymorphisms associated with resistance to artemisinins

The study protocol obtained ethical clearance by the National Bioethics Committee for Health of Mozambique (CNBS—IRB00002657) (Ref: 131/CNBS/2021) dated: March 2021.

Malaria is endemic throughout Mozambique, ranging from hyper-endemic areas along the coastline, meso-endemic areas in the interior lowlands and some hypo-endemic areas in the interior highlands. Several factors contribute to this endemicity, ranging from climatic and environmental conditions, such as favourable temperatures and rainfall, as well as favourable breeding sites for the vector. Most of the country has year-round throughout the year, with peaks during the rainy season from December to April [4]. Patient recruitment took place in 3 different epidemiological settings, in northern (Hospital distrital de Marrupa in Niassa), central (Centro de Saúde Eduardo Mondlhane and Centro de Saúde 7 de abril in Manica) and southern (Hospital provincial da Matola in Maputo) area of Mozambique, between April and August of 2021. One hundred and fifty samples were collected in each study site (Fig. 1A). The choice of Niassa, Manica and Maputo provinces not only is aligned with the study scientific goals but also with the objectives of Instituto Nacional de Saúde, and National Malaria Control Programme Mozambique, contributing to the systematic mapping of malaria cases at the national level.

A total of 450 participants of all ages with malaria positive Rapid Diagnostic Test (RDT) were recruited and provided 100 μL of blood samples on filter papers (Whatman® FTA® cards), after written informed consent. All dried blood spot samples were then stored under − 20 °C until they were used for genotyping.

Parasite genomic DNA from dried blood spots was extracted using the Chelex method [32], and DNA was stored at − 20 °C. Real-time PCR was used for P. falciparum confirmation. PCR reactions targeting the 18S rRNA gene were conducted as described in Rosanas-Urgell et al. 2010, with modifications. Briefly, forward primer 5′-TATTGCTTTTGAGAGGTTTTGTTACTTTG-3′ and reverse primer ACCTCTGACATCTGAATACGAATGC and the probe FAM-ACGGGTAGTCATGATTGAGTT-MGB-BHQ were used. PCR reaction mixture consisted of 7.5 μL of 2X (NZYTECH, Portugal), 600 nM of each primer and 200 nM of FAM™-labeled probe (IDT Integrated DNA Technologies, USA), 1 μL of genomic DNA and water up to 15 μL. PCR conditions: 50 °C for 2 min and 95 °C for 10 min; these were followed by 40 cycles at 94 °C for 30 s and a final cycle at 60 °C for 1 min. Triplicate samples were assayed in the Bio-Rad 500 Real Time PCR System™ (Applied Biosystems, USA). All reactions were performed with positive controls (DNA from 3D7 strain of P. falciparum culture).

The Pfk13 fragment, containing the main polymorphisms associated with resistance to artemisinin, was amplified by nested PCR as described by Escobar and co-workers [29], with slight adjustements. Briefly, specific primers were developed for this purpose (forward—5′-CTATACCCATACCAAAAGATTTAAGTG-3′, reverse—5′-GCTTGGCCCATCTTTATTAGTTCCC-3′), obtaining a fragment of 902 bp (from codon 412 to codon 723). PCR conditions [33]: 94 °C 3 min; [94 °C 30 s, 57 °C 30 s, 72 °C 30 s] 10×; [94 °C 30 s, 55 °C 30 s, 72 °C 30 s] 30×; 72 °C 3 min. PCR products were analysed by electrophoresis on a 2% agarose gel stained with GreenSafe Premium (Nzytech, Portugal) to confirm amplification of targeted fragments. All positive PCR products were then purified using SureClean Plus (Bioline) and shipped to Eurofins Genomics (GATC services, Germany), to proceed with Sanger sequencing. Successfully sequenced samples were aligned to the PF3D7_1343700 gene using Multalign software (http://​multalin.​toulouse.​inra.​fr/​; free online) and/or BioEdit version 7.2 for mutation detection. Bar charts with prevalence of artemisinin resistance markers were generated using GraPhpad Prism 8.01 software. In order to predict the potential impact of non-synonymous SNPs on protein function, we used SIFT software (Sorting Intolerant From Tolerant, free online, https://​sift.​bii.​a-star.​edu.​sg/​index.​html). SIFT software, takes into account the position at which the variation takes place and the type of amino acid change, then, chooses related proteins and obtains an alignment of these proteins with the query [34]. Finally, it calculates the probability that this particular amino acid change will is tolerated [34]. If the calculated value is less than a cutoff of 0.05, the substitution is predicted to be deleterious, and the opposite is considered not deleterious [34‐36].

From the 450 participants included in the study, one third were from each of the three study provinces (Fig. 1). Most of the participants were males, representing 57, 53% (259/450). The age ranged from 6 months to 74 years, and the overall majority (%) were children between 1 to 12 years old, and the average was 15 years old.

From selected samples, 66, 2% (298/450) were confirmed for P. falciparum by real time PCR, and DNA was successfully sequenced for pfk13, where 98, 100 and 100 were from Niassa, Manica and Maputo, respectively (Table 1). No validated SNP for artemisinin derivate resistance was detected in our samples. However, in Niassa, 10.2% (9/98) of samples harbored nine different non-synonymous mutations (Fig. 1B, Table 1). In Manica and Maputo, five different non-synonymous mutations were detected in 6 (6/100) and 5 (5/100) samples, respectively (Fig. 1C, D). 56.3% (9/16) of the non-synonymous mutations reported in this study were by substitution at the 1st base of the codon, 25% (4/16) at the second base and 18.8% (3/16) at the third base. The non-synonymous mutation, W470R, was detected in all provinces and its prevalence decreases from north to south with 3%, 2% and 1% in Niassa, Manica and Maputo, respectively. Synonymous mutations were only detected in samples from Manica and Maputo (Fig. 1E).

With a cut-off score of 0.05, we predicted that V454E, V494L, A578P, V581A, E509D, E455Q and L663I mutations would be tolerated or have a non-deleterious effect on protein function, with equal score or higher than 0.05 as shown in the Fig. 1E. For the remaining SNPs, were predicted to affect protein function [34‐36] with a score of 0.00 (Fig. 1E).