Background

Artemisinin-based combination therapies (ACTs) are the mainstay of treatment for uncomplicated malaria. Uganda has used artemether-lumefantrine (AL) as the first-line treatment for uncomplicated malaria since 2006. However, recent observations of both artemisinin partial resistance and reduced lumefantrine susceptibility raise concerns about the potential for decreased efficacy of both components of the combination therapy.

Existing surveillance efforts have primarily focused on the Kelch13 (K13) mutation, multidrug resistance protein 1 (MDR1), and chloroquine resistance transporter (CRT), which are associated with artemisinin resistance. While these markers partially explain changes in lumefantrine response, their effects are limited, and a definitive resistance determinant has not been identified. The researchers moved beyond examining only specific candidate genes and performed whole-genome sequencing of clinical isolates from Uganda to track regions that have undergone recent natural selection.

Key Findings

The researchers analyzed whole-genome sequences of 157 Plasmodium falciparum isolates obtained from patients with uncomplicated malaria. 88% of the isolates had an average sequencing depth of 50x or greater, and the analysis included 55,100 high-quality single nucleotide polymorphisms (SNPs) with a minimum allele frequency of 2%. Calculation of the extended haplotype homozygosity (EHH) score and the integrated haplotype score (iHS) for each allele revealed that the strongest selection signal was not associated with K13, but rather with a region of approximately 260 kilobases on chromosome 7.

The signal was centered on the px1 gene, which encodes phosphoinositide-binding protein PX1. The researchers identified a haplotype consisting of the L1222P, M1701I, and D1705N amino acid substitutions, as well as two in-frame deletions of 18 and 36 nucleotides, which they named 'PIN'. This haplotype was not detected in samples from 2004, but first appeared in 2008, and its frequency reached 55% in eastern Uganda and 84% in northern Uganda in 2024. Among the 1,598 samples used for spatiotemporal analysis, 1,436 were sequenced to a depth of 25x or greater.

In K13 wild-type parasites, the median half-maximal inhibitory concentration (IC50) for lumefantrine in the PIN-bearing group was 14.3 nanomol, 2.3-fold higher than that of the wild-type LMD haplotype (6.2 nanomol). For mefloquine, the IC50 values were 17.1 and 11.0 nanomol, respectively, and for dihydroartemisinin (DHA), the values were 3.7 and 1.8 nanomol, respectively. All three drugs showed statistically significant reductions in susceptibility. Two clones of the 3D7 parasite in which px1 was disrupted were found to be more sensitive to all three drugs in a 72-hour assay, providing functional evidence that PX1 regulates drug response.

Implications and Outlook

The PIN haplotype may be involved in susceptibility to both the artemisinin component and the lumefantrine component of AL. The fact that it emerged 8 years before the first identification of K13 mutations in 2016, and that it has independent effects on K13 and MDR1, is noteworthy. The long-term use of AL in Uganda may have first selected for PIN, followed by the emergence of K13 partial resistance mutations, which is a plausible evolutionary scenario.

However, reduced susceptibility does not necessarily equate to clinical resistance. There was no significant difference in ring-stage survival between the PIN and LMD haplotypes, and there is currently limited evidence that changes in in vitro IC50 values directly predict treatment failure. The region with the selection signal contains 69 genes, and it is difficult to isolate the effects of individual substitutions and deletions in PIN using only gene knockout experiments. Future work should include prospective studies that link patient outcomes and genomic surveillance in neighboring African countries.