Drug Resistance Molecular Markers of Plasmodium falciparum and Severity of Malaria in Febrile Children in the Sentinel Site for Malaria Surveillance of Melen in Gabon: Additional Data from the Plasmodium Diversity Network African Network

The objective of this study was to analyze the relationship between the frequency of artemisinin-based combination (ACT) drug resistance molecular markers and clinical forms of P. falciparum malaria and parasitemia. A cross-sectional study was carried out between January and April 2014 at the Operational Clinical Research Unit of Melen in febrile children aged 12 to 240 months with a Plasmodium sp. infection. A total of 3 mL of peripheral blood collected from an EDTA tube was used for leukocyte depletion. DNA mutation detection was performed by next generation sequencing (NGS). A total of 1075 patients were screened for malaria. Among them, 384 had a Plasmodium infection. P. falciparum mono-infection was found in 98.9% of the patients. Pfcrt-326T mutation was found in all isolates, while 37.9% had Pfmdr2-484I mutant allele. The highest median parasite densities were found in patients infected by parasites carrying the CVIET haplotype of the Pfcrt gene. The different genetic profiles found here, and their variations according to clinical and biological signs of severe malaria, are additional arguments for the surveillance of P. falciparum strains.


Introduction
Malaria cases were estimated at 247 million in the world in 2021. During this year, the number of malaria deaths was 619,000, of which 80% were children under the age of 5 [1]. Plasmodium (P.) falciparum is responsible for a severe form of the disease. Since the start of the 2000s, the World Health Organization (WHO) has recommended the use of sulfadoxine-pyrimethamine combination for preventive treatment against malaria during pregnancy, and artemisinin-based combinations (ACTs) for the treatment of uncomplicated malaria. The main combinations are artemether-lumefantrine (AL), artesunateamodiaquine (AS-AQ), dihydroartemisinin piperaquine (DH-PQ), artesunate-mefloquine (AS-MQ) and artesunate-sulfadoxine-pyrimethamine (AS-SP). Severe malaria is treated with injectable artesunate (intramuscular or intravenous), followed by a complete three-day ACT course once the patient can tolerate oral medicines. Artemether and quinine may also be used in the absence of artesunate [2].
These recommendations have been adopted after the spread of P. falciparum strains multiresistant to conventional antimalarials, some of which are partner's molecules of artemisinin derivatives in ACTs. ACTs led to a significant reduction in malaria morbidity and mortality in populations at risk [1,3]. Thus, WHO recommends regular monitoring of the efficacy of these antimalarial medicines to ensure that the chosen treatments are still efficacious.
In Southeast Asia, the resistance of parasites against artemisinin (ART) derivatives, as well as Piperaquine, has been shown [4]. Resistance to ART and its derivatives is characterized by an extension of the parasite clearance time in treated individuals [4,5]. Indeed, the persistence of parasites after three days of ACT treatment has been observed [5].
Other mutations have been found in other genes, such as P. falciparum ferrodoxine (Pffd-D193Y), P. falciparum apicoplast ribosomal precursor S10 (Pfarps-V127M), P. falciparum multidrug resistance protein 2 (Pfmdr2-T484I) and P. falciparum chloroquine resistance transporter (Pfcrt-I356T and Pfcrt-N326S) [11,12]. These mutations may contribute to the development of the parasite's resistance to ART [13]. Likewise, the T484I mutation of P. falciparum multidrug resistance protein 2 (Pfmdr2) has been associated with artemisinin drug resistance [10,13]. Such resistance may be related to a higher risk of developing severe malaria linked to a change in the fitness of the P. falciparum strains.
Monitoring the resistance of malaria parasites to partner drugs is also necessary [14]. Two genes, Pfcrt and Pfmdr-1, are associated with resistance to the three main partner drugs: amodiaquine (AQ), lumefantrine and mefloquine. The 74I-75E-76T mutations of the Pfcrt gene and those of the Pfmdr1 gene, N86Y-F184Y-D1246Y, are related to the parasite's resistance to chloroquine (CQ) and AQ [15,16]. The N86-F184-D1246 haplotype of the Pfmdr1 gene and the mutation detected at the E415G codon of the gene (exo-E415G) coding for exonuclease are associated with a decrease in parasite susceptibility to lumefantrine and piperaquine, respectively [17,18].
In Gabon, the treatment of uncomplicated malaria is based on the administration of AS-AQ and AL as first-line treatment and dihydroartemisinin-piperaquine phosphate (DHA-PQ) as second-line treatment. The injectable form of artesunate or quinine (if artesunate is unavailable) is recommended for the treatment of complicated malaria. Studies performed in different areas of this country reported a high frequency of molecular markers associated with resistance to aminoquinoline antimalarials and antifolates [19,20]. Otherwise, malaria prevalence in Gabon, after a drop between 2005 and 2011, has increased again in patients aged less than 20 years: 25% in 2011 and 36.3% in 2014 [21]. The prevalence of P. falciparum malaria was 32.8% among 1962 patients received in 2021 in the sentinel site for malaria surveillance of Melen (unpublished data). Children under 5 years old and those over 5 years old accounted for 68.9% and 54.5%, respectively (unpublished data). Among those aged over 5 years, 45.4% concerned children aged over 11 years (unpublished data). Likewise, the prevalence of severe malaria increased and a change in the clinical-biological profile was observed, with neurological forms being predominant [22].
Thus, the aim of this study was to analyze the relationship between the frequency of drug resistance molecular markers to ACTs and clinical forms of P. falciparum malaria according to parasite density level.

Study Area
A cross-sectional study was conducted between January and April 2014 in the Operational and Clinical Research Unit (OCRU), located in the Regional Hospital of Melen (RHM) in the north of Libreville. The RHM is located 11 km north of Libreville, as previously described [21]. Its equatorial climate is subdivided in four seasons: a short dry season from December to January, a long dry season from May to September, a short rainy season from October to November and a long rainy season from February to June. The malaria transmission is perennial in the country. P. falciparum is the cause of 99% of the symptomatic infections [20]. The major vector species are Anopheles (A.) gambiae and A. funestus [23].

Study Population
The participants were children aged 12 to 240 months mainly living in the district surrounding the RHM. Children meeting the following criteria were invited to participate: being febrile (with a tympanic temperature of >37.5 • C) or having a 24 h to 48 h history of fever before the day of consultation. A clinical examination was performed to identify the criteria of complicated malaria, looking for repeated convulsions, iterative vomiting and icterus in particular. Complicated malaria was defined and classified according to WHO indications [24]. Biological analyses were also performed for each patient. Indeed, health workers appointed in pediatric wards directed all patients towards the OCRU to benefit from the blood smear and cell blood count tests. Those who had a P. falciparum mono infection and parasite density of >2000 trophozoïtes per microliter of blood (p/µL) were included. Non-inclusion criteria were the presence of a mixed Plasmodium infection and non-falciparum infection.
Approximately 3 mL of blood sample in a EDTA tube were used for the malaria diagnoses.

Malaria Diagnosis
Giemsa-stained thick and thin blood smears of peripheral blood were taken for each patient, as previously described [25]. A smear was only considered negative if no malaria parasites were seen in 100 microscopic fields using immersion oil. The thin blood smear made it possible to identify Plasmodium species. All patients infected with Plasmodium species were treated according to national recommendations.

Leukocyte Depletion
Leukocyte depletion consists of the elimination of blood leukocytes before malaria parasite DNA extraction using CF11 columns containing cellulose powder. It reduces the amount of human DNA (contained in the nucleus of leukocytes) in a Plasmodium infected blood sample to obtain P. falciparum purified DNA. Leucocyte depletion was carried out at the Department of Parasitology-Mycology of the Faculty of Medicine. Venous blood sample from each patient was depleted of leucocytes within 6 hours following the sample collection. After the depletion of the leucocytes, a quality control of the sample was done based on microscopy.

Molecular Analysis
Parasite genomic DNA was then extracted from samples after leucocytes depletion using a commercial Kit QIAamp DNA Blood Midi (100), according to the manufacturer's instructions. DNA was kept in 100 µL of elution buffer at −80 • C for long-term conservation. The genes analyzed are reported in Table 1.
The NGS was carried out at the Wellcome Trust Sanger Institute (United Kingdom) for the detection of mutations related to resistance to ART and quinolines. To characterize ART, the resistance genes designated parasite genetic background (PGB) were sequenced (Table 1). Polymorphism on the Pfcrt (at codons 72-76) and Pfmdr1 (at codons 86, 184 and 1246) genes was also analyzed in order to characterize the genetic profile of Plasmodium strains resistant to quinolines. All single nucleotide polymorphisms (SNPs) were further assessed using the 3D7 strain as a reference genome. This work was carried out as part of a multicenter study by different teams in 10 African countries for PfK13 gene polymorphism analysis [30]. These teams gathered in the "Plasmodium Diversity Network African (PDNA)" network, aiming to study the genetic diversity of P. falciparum. This includes the study of molecular markers associated with resistance to artemisinin and partner molecules in P. falciparum isolates [26]. This network has established a collaboration with the Wellcome Trust Sanger Institute in order to generate standardized data.

Genetic Report Card User Guide
The Genetic Report Card (GRC) contains malaria parasite genetic data, derived from the analysis of patient blood samples. For each sample, we determined the genotype for several variations (single nucleotide polymorphisms, or SNPs) that are known to be associated with resistance to antimalarials. The report also contains data on Plasmodium species found in the sample. The following is a guide to interpreting the Genetic Report Card data, linking the genotyped mutations (molecular markers) with resistance to different antimalarial drugs (Tables A1 and A2).

Drug Resistance Genes and Haplotypes
Several columns in the GRC show the genotypes of drug resistance mutations. For consistency with the published literature, these genotypes are reported as amino acid positions and alleles. For genes that carry multiple SNPs of interest, we concatenated the genotypes to form a haplotype, which is also common in the literature. In the haplotypes, we also used the following special characters: A dash ("-") indicates a missing genotype, either because the sample could not be genotyped, or because the assay is still under development/refinement. An asterisk ("*") indicates that two alleles were detected (a heterozygous call) (Table A2).

Sample Barcodes
The sample barcodes are formed by concatenated genotypes at 101 SNPs across the P. falciparum genome. These SNPs are all biallelic, i.e., only two alleles are observed. They were chosen for their usefulness in analyses of relationship between parasites, and are not associated with drug resistance. The full list of SNPs used is given in Appendix A.
SNPs within the barcode are represented by the observed nucleotide (A, T, C and G). If the genotype is missing (could not be detected), the symbol "X" is used; the symbol "N" indicates that both alleles were observed (heterozygous call).

Species Co-Infections
We detected the presence of different Plasmodium species by testing an SNP in the parasite mitochondria (PfMIT: 270) that carries different alleles in three different parasite species: P. falciparum, P. vivax and P. knowlesi. This assay has not yet been published but has been validated with quantitative PCR assays for detecting P. vivax and P. falciparum.

Artemisinin Drug Resistance
Parasite genetic background (PGB) mutations showed evidence of a genetic background of mutations that allowed for the emergence of PfK13 mutations [11]. These mutations are: V127M in the Pfarp-s10 (PF3D7_1460900) protein, D193Y in ferredoxin (PF3D7_1318100), N326S and I356T in Pfcrt (PF3D7_0709000) and T484I in Pfmdr2 (PF3D7_1447900). These are displayed in a concatenated haplotype form, with the reference allele (WT) being VDNIT.

Chloroquine Drug Resistance
Chloroquine drug resistance is primarily mediated by mutations in the chloroquine resistance transporter (Pfcrt, PF3D7_0709000) [31]. An accessory mutation at position 86 in the multidrug resistance protein (Pfmdr1, PF3D7_0523000) has been shown to accentuate this resistance phenotype in parasites [32]. The loci in Pfcrt are represented as a 5-amino acid haplotype at positions 72-76, with the wild-type haplotype being CVMNK [29]. The CVIET haplotype is the most widespread resistant haplotype in Asia and Africa, while SVMNT is common in resistant parasites in South America and Oceania. The Pfmdr1 mutation at position 86 is the first in the three amino-acid haplotypes reported for this gene. The WT variant is N, while the Y variant in MDR-1 86 enhances resistance to chloroquine.

Amodiaquine Drug Resistance
Mutations in the multidrug resistance protein (Pfmdr1, PF3D7_0523000) have been associated with parasite response to amodiaquine. There is limited evidence that mutations at positions 86 and 1246 can mediate the response to this drug, and these positions constitute the haplotype reported for this gene [33]. In vitro experiments have shown that haplotypes containing the mutant 86Y have increased IC50's to chloroquine and amodiaquine.

Mefloquine and Lumefantrine Drug Resistance
Mutations in the multidrug resistance protein (Pfmdr1, PF3D7_0523000) have been associated with parasite response to the drugs mefloquine and lumefantrine [28]. There is limited evidence that variations at positions 86, 184 and 1246 increase susceptibility to these drugs [28].

Piperaquine Drug Resistance
In a recent genome-wide association study (GWAS), an SNP in a putative exonuclease gene (PF3D7_1362500) was associated with ex vivo piperaquine IC50 of parasite isolates from Cambodia [17]. This molecular marker is at position 415, and a G allele was shown to be significantly associated with increased tolerance of piperaquine with respect to the WT allele E.

Ethical Considerations
In Gabon, the Gabonese Ministry of Health, represented by the Malaria National Control Program, has committed the Department of Parasitology-Mycology and Tropical Medicine of the Faculty of Medicine of Libreville to perform an evaluation of malaria prevalence and antimalarial drug resistance monitoring in sentinel sites throughout the country. In this study, all patients were enrolled in the sentinel site for malaria surveillance in the RHM (Melen). In this context, patients and the guardian or parent of the children included were informed of the activities in this sentinel site and their signed consent was obtained. Children over the age of 12 gave their consent to participate in the study. In addition, in this site, a standardized procedure on the conservation of blood aliquots of febrile patients in the biobank was established. Being in a sentinel site, the participants were treated according to national recommendations.

Statistical Analysis
Data were analyzed using Epi InfoTM7. All variables were compared using Chi 2 or the Fisher test to compare the exact proportions, and Student's t-test and analysis of variance (ANOVA) or the Kruskal-Wallis test, as appropriate for continuous variables. p-value < 0.05 was considered statistically significant.

Characteristics of the Study Population
A total of 1074 patients were screened for malaria. Median age of the population was 69 (66-72) months, with a sex ratio of 1.23 (with 55.1% of patients of male sex). Among them, 384 had a Plasmodium infection. P. falciparum mono-infection was found in 98.9% of blood samples. Among those, sixty-one samples were selected for molecular studies (Figure 1).
Medicine of the Faculty of Medicine of Libreville to perform an evaluation of malar prevalence and antimalarial drug resistance monitoring in sentinel sites throughout th country. In this study, all patients were enrolled in the sentinel site for malar surveillance in the RHM (Melen). In this context, patients and the guardian or parent the children included were informed of the activities in this sentinel site and their signe consent was obtained. Children over the age of 12 gave their consent to participate in th study. In addition, in this site, a standardized procedure on the conservation of bloo aliquots of febrile patients in the biobank was established. Being in a sentinel site, th participants were treated according to national recommendations.

Statistical Analysis
Data were analyzed using Epi InfoTM7. All variables were compared using Chi 2 the Fisher test to compare the exact proportions, and Student's t-test and analysis variance (ANOVA) or the Kruskal-Wallis test, as appropriate for continuous variables. value < 0.05 was considered statistically significant.

Characteristics of the Study Population
A total of 1074 patients were screened for malaria. Median age of the population w 69 (66-72) months, with a sex ratio of 1.23 (with 55.1% of patients of male sex). Amon them, 384 had a Plasmodium infection. P. falciparum mono-infection was found in 98.9% blood samples. Among those, sixty-one samples were selected for molecular studi (Figure 1). Amplification and sequencing of the Pfcrt, Pfmdr2, Pfarps10, Pffd and Pfmdr1 genes were successfully performed in 58 samples. These samples were from febrile patients, of whom 70.7% (n = 41/58) had uncomplicated malaria (UM) ( Table 2). The median parasitemia was 32,200 (25,200) p/µL in clinical samples. The highest median parasite density was found in samples from patients with complicated malaria (CM) (p < 0.01) ( Table 2). Hyperparasitemia (parasite density ≥200,000 p/µL) was found in six samples from patients with CM. Likewise, diarrhea, vomiting and convulsions were observed in two, three and six patients with CM, respectively. The use of antimalarial drugs before consultation applied to 17.6% of patients with CM and 82.3% of those with UM.

Frequency of Haplotypes According to Malaria Clinical Forms
Considering the different haplotypes identified, the double mutant VDNTI haplotype was 1.4-fold (35.3% vs 24.4%) more frequent in isolates from patients with CM, although the difference was not statistically significant (p = 0.3). Similarly, wild type haplotype NYD was almost twice as frequent in the isolates of these patients (23.5%; n = 4/17 vs 12%; n = 5/41 in isolates of patients with uncomplicated malaria) (p = 0.4; OR 2.2154 [0.5146-9.5372]).

Frequency of Haplotypes According to Malaria Clinical Forms
Considering the different haplotypes identified, the double mutant VDNTI haplotype was 1.4-fold (35.3% vs. 24.4%) more frequent in isolates from patients with CM, although the difference was not statistically significant (p = 0.3). Similarly, wild type haplotype NYD was almost twice as frequent in the isolates of these patients (23.5%; n = 4/17 vs. 12%; n = 5/41 in isolates of patients with uncomplicated malaria) (p = 0.4; OR 2.2154 [0.5146-9.5372]).

Haplotypes and Clinical and Biological Signs of Severe Malaria
Patients with hyperparasitemia or hyperparasitemia associated with convulsion were frequently infected by parasites with the CVIET triple mutant haplotype, VDNTT haplotype, NFD and YFD haplotypes of Pfmdr1 gene. The VDNTI haplotype was frequent in isolates from patients with icterus and convulsions.
The combination of PGB and Pfcrt 72-76 haplotypes showed the presence of five haplotypes (H0 to H4) ( Table 3). H0 and H2 were the most frequent (Table 3). Table 3. Genetic profile of P. falciparum isolates according to median parasite density.

Discussion
In the present study, the monitoring of drug resistance molecular markers related to antimalarial drugs used in Gabon revealed mutations T484I (>37.8%) and I356T (100%) on the Pfmdr2 and Pfcrt genes, respectively. In Asia, these are associated with a delayed clearance of parasites in patients treated with AL, DHA-PQ and AS-AQ, respectively [13]. The appearance of these mutations suggests that P. falciparum strains are under drug pressure. Thus, the occurrence of additional mutations on the Pfarps and Pffd genes may be expected, although none were found in these isolates. Similarly, molecular markers such as Exo-E415G SNP related to piperaquine resistance and DH-PQ treatment failure were not detected in the isolates. Nevertheless, P. falciparum strains carrying multiple copies of the Plasmepsin 2 gene have been collected in previous studies carried out in Gabon, Burkina Faso, the Democratic Republic of Congo, Mozambique and Uganda [34]. Taken altogether, these data suggest a delayed appearance of the Exo-E415G mutation, compared to the Plasmepsin 2 gene's amplification in isolates circulating in areas where these molecules are used. Furthermore, the mutant Pfmdr1-86Y (43%) allele, which is less frequent than the mutant Pfcrt-76T allele (60%) in P. falciparum isolates from Melen was also detected. Its frequency confirms a decreasing frequency of this haplotype since the implementation of ACTs [19,35].
It is notable that multiple mutations are generally required to functionally confer and increase the level of drug resistance. The Pfcrt CVIET triple mutation was found in 60.4% of P. falciparum isolates from Melen. This proportion is similar to those reported in the Congo (71%) and south-east Gabon (70.6%) [36,37]. This haplotype was reported to be more frequently detected in isolates from West Africa (44%) and East Africa (38%), than in those from South Africa (21%) and Central Africa (26%) [38]. Among the Pfmdr1 haplotypes, the NFD haplotype related to lumefantrine resistance was the most prevalent, found in almost one out of two isolates, several years after the introduction of AL and AS-AQ in Gabon. Similar frequencies, varying from 35% to 46%, were found in other areas of Gabon and in Tanzania [20,36]. In contrast, a higher prevalence of the NFD haplotype was found in Senegal (62.26%), Kenya (51%) and Ghana (60%) [38]. The circulation of P. falciparum strains carrying the NFD haplotype is frequently related to pressure due to the use of AL [20,39].
In addition to the use of antimalarial drugs as self-medication, as well as the nonobservance of the recommended dosage, other factors, such as the parasite's genetic background, may lead to an increase in the number of complicated malaria cases in Gabon [22].
In the present study, specific haplotypes varied according to malaria severity, as well as clinical and biological signs. Haplotypes NFD, YFD and CVMNK were more common in isolates from patients with uncomplicated malaria. In contrast, in those from patients with severe malaria, NYD and VDNTI were the most frequent, although these differences were not statistically significant. The CVIET haplotype was common in both groups of patients. The relationship between these haplotypes and the occurrence of severe malaria needs to be confirmed. Specific haplotypes have been identified, but no significant relationship between the clinical forms of malaria-or biological modifications and the frequency of drug resistance molecular markers-has been found. More specifically, patients with hyperparasitemia were infected by parasites with the CVIET and VDNTT haplotypes. The VDNTI haplotype was found in isolates from patients with icterus and convulsion. NFD and YFD were detected in parasites from patients with convulsion. A relationship between NYD and high parasitemia (p = 0.04) has been reported elsewhere [40]. In the current study, the combined haplotypes H2 (VDNTT + CVIET), H3 (VDNTT + CVIET/CVMNK) and H4 (VDNTT + CVIET/CVMNK) were found in patients infected with high parasite densities, suggesting their involvement in the parasite's ability to multiply.

Conclusions
More than one decade after the withdrawal of CQ and the introduction of ACTs in Gabon, monitoring of drug resistance molecular markers remains critical to providing data on the propagation of artemisinin resistant parasites, as done in the present study. Moreover, the different genetic profiles found here and their variations according to the clinical and biological signs of severe malaria are additional arguments for their monitoring, and for an in-depth analysis of their role of the CIVIET haplotype of the Pfcrt gene on the fitness of the parasite.

Patents
This section is not mandatory but may be added if there are patents resulting from the work reported in this manuscript.