Background
Malaria caused by
Plasmodium falciparum infection is a significant public health problem in Nigeria where over 30% of the global burden of the disease is borne despite recent mass scale-up of intervention measures [
1]. Artemisinin-based combination therapy (ACT) is recommended as first line treatment against uncomplicated malaria and it remains a critical tool in malaria control programmes. The most commonly used ACT, Artemether-Lumefantrine (AL), remains largely efficacious in sub-Saharan Africa [
1]. The efficacy of AL has been attributed to the fast action of artemisinin and the resultant rapid reduction in the density of asexual malaria parasites, while lumefantrine acts to prevent recrudescence [
2]. However, resistance to artemisinin-based drugs has been reported in many endemic regions in South-East Asia [
3]. Although evidence of resistance among African populations of
Plasmodium falciparum remains scarce, there have been emerging cases of potentially adaptive strains and slow parasite clearance following treatment [
4,
5]. These parasites are often below the threshold of detection and are rarely re-treated, thus contributing to disease transmission and spread of resistance [
6]. Accurate detection, quantification and characterization of such parasite isolates after treatment is necessary for the proper evaluation and monitoring of drug efficacy.
One recent technique for the ultra-sensitive detection of low parasitaemia involves the amplification of the
var gene family present in the sub-telomere of the parasite [
7]. Each parasite isolate comprises about 50–150
var genes [
8], which possess acidic terminal sequence (ATS) with well-conserved domains that are targeted. Once detected, parasite isolates before and after treatment can be genetically fingerprinted with a single nucleotide polymorphism (SNP) barcode, which is another sensitive assay that requires only a small amount of input DNA and can be used to distinguish recrudescence from re-infection [
9]. Prompt detection of mutations in loci implicated in drug resistance such as multidrug resistance gene (
mdr1), chloroquine resistance transporter (
crt), dihydrofolate reductase (
dhfr), dihydropteroate synthase
(dhps) and kelch-13 (
K13) genes provides molecular evidence of antimalarial resistant parasites [
10,
11]. For accurate surveillance of ACT resistance in a malaria-endemic setting as Nigeria, there is a need for continuous therapeutic efficacy studies using sensitive techniques that enable early detection of tolerant parasite strains. In this study, we assessed the efficacy of artemether-lumefantrine, determined genetic relatedness of parasites isolated before and after treatment and the prevalence of markers of drug resistance in the parasite population.
Discussion
One of the key approaches to malaria control is prompt treatment of patients for an improvement in health and prevention of onward transmission. With rising reports of drug-resistant Plasmodium falciparum, surveillance of ACT efficacy becomes important so that recrudescent infections can be promptly detected, treated and documented. In this study, we reported in vivo re-appearance of parasites following appropriate treatment of uncomplicated P. falciparum with artemether-lumefantrine in Nigeria. The genomic barcode analysis of pre-treatment and re-appearing infections suggested recrudescence despite the absence of mutation in K-13 C580Y. This suggests that the potency of the most promising antimalarial drug, artemisinin-based combination, in treating P. falciparum infection in Nigeria may be progressively waning.
High genetic diversity of the parasites was observed in the population. This observation is in line with expectations from high transmission settings where multiple parasite genomes in each sample and highly variable combinations of barcodes among individuals in the population have been observed [
19]. We also observed a marginal increase in CoI 28 days after treatment suggesting a possible lack of clonal expansion following treatment. This agrees with the observation in Kenya where, contrary to expectations of reduced CoIs caused by the persistence of only drug-resistant strains in post-treatment infections, CoI on day 28 was higher than day 0 [
20].
Age and immunity have previously been demonstrated as indicators of varying therapeutic responses [
21]. Thus, older individuals living in a high transmission area who have a more developed acquired immunity than children in the same setting would be expected to have fewer persistent infections following treatment. While our finding could not directly correlate age with the persistence of parasitaemia on the third day of treatment, it revealed re-appearance of infections on day 28 only among children less than fifteen years old. This observation possibly explains the role of acquired immunity in complementing antimalarial drug action in parasite clearance. However, additional studies are required to understand the influence of immunity on therapeutic response and drug resistance by correlating parasite clearance rates with malaria antibody titres in the individuals.
Although we did not observe any clinical outcome of persistent infections and there was no direct association between re-appearance of infection and day 0 parasite densities, it is important to continue monitoring the prevalence of residual infections following treatment, even though they were observed in low quantities in this study. Since parasitaemia as low as 0.5/μl of blood corresponds to as high as approximately one million parasites in the body [
21], low-grade infections observed after treatment in this study area may sustain a reservoir for silent transmission of the disease. Studies that involve longer follow-up durations and laboratory adaptation of the sub-microscopic parasites may provide more information on the transmission potential of these post-treatment parasites.
A high frequency of mutant alleles of
crt and
mdr1 (N86Y and Y134F) was reported in this study, but these mutations have not been directly associated with post- treatment re-appearance of infections. In addition, contrary to a previous report showing that polymorphisms in
mdr1 genes were associated with decreased sensitivity to lumefantrine [
22], our phenotype-genotype association study provided no evidence that the presence of 86Y allele sensitized parasites to lumefantrine as there was no correlation between the distribution of N86 and 134F alleles and presence of parasites on day 28. An increase in the frequency of
Pfcrt 76 T is associated with reduced sensitivity to chloroquine [
23] and the removal of drug pressure is expected to cause a decline in resistance conferring mutations [
24]. The prevalence of mutant 76 T allele in this present investigation is lower than previously reported before the proscription of CQ for mild malaria treatment in Nigeria [
25]. This may be a pointer to ongoing selection for chloroquine-susceptible parasites following widespread use of artemisinin. Kelch-13 sequencing information from recent therapeutic efficacy studies conducted in Cambodia, Lao and Vietnam has revealed cysteine to tyrosine mutation in codon 580 of K13 gene (C580Y) as the dominant polymorphism associated with artemisinin resistance [
3]. However, our investigation revealed the absence of this mutation in re-appearing parasites.
We propose further investigations involving complete genome-wide associations between pre-treatment and persistent parasites to circumvent the limitation of our study which did not preclude the influence of re-infection with the same parasite type or residual parasite DNA and non-replicating mature gametocytes that can persist at low concentrations after treatment. Computational analyses that involve the combination of genomic information with epidemiological simulations will present complementary tools for interpreting population-level impact of drug interventions. In addition, since
P. falciparum is thought to exhibit drug-induced developmental arrest until the activity of artemisinin in the host has diminished [
26], details about quiescence in post-treatment infections using in vitro models may further explain the underlying mechanisms of re-appearance of parasites after treatment in the study area. A broader line of investigation on the efficacy of ACTs in larger populations in the country and across the borders is necessary to understand spatial and temporal heterogeneity of infectious reservoirs post-treatment.
Acknowledgements
The authors appreciate the co-operation of participants and the efforts of the staff and management of the health facilities visited. Special recognition goes to Dr. S. Babatimilehin for coordinating the patients on follow-up. We thank the Department of Immunology and Infectious Diseases, Harvard TH Chan School of Public Health, Boston, MA, USA for the training on Molecular SNP Barcoding and HRM analysis. The authors acknowledge Dr. Rachel Daniels for her comments on the manuscript.