The extended recovery ring-stage survival assay is a scalable alternative for artemisinin susceptibility phenotyping of fresh Plasmodium falciparum isolates

ABSTRACT

Artemisinin partial resistance (ART-R) has emerged in eastern Africa, necessitating regular surveillance of susceptibility of Plasmodium falciparum to artemisinins. The microscopy-based ring-stage survival assay (RSA) provides a laboratory correlate of ART-R but is limited by low throughput and subjectivity of microscopic counts of viable parasites. The extended recovery ring-stage survival assay (eRRSA) replaces microscopy with efficient quantitative PCR (qPCR) readouts but has been studied only with culture-adapted P. falciparum clones. We measured susceptibility to dihydroartemisinin (DHA) after a 6-h incubation with 700-nM DHA, followed by culture without drug, by comparing survival with that of untreated controls by microscopy (the RSA) or qPCR (the eRRSA) and also performed standard growth inhibition (half-maximal inhibitory concentration [IC₅₀]) assays for 122 P. falciparum isolates freshly collected in eastern and northern Uganda from March to July 2022. The median values for RSA survival, eRRSA fold change, and DHA IC₅₀ were 3.0%, 46.2, and 3.2 nM, respectively. RSA percent survival and eRRSA fold changes correlated strongly (Spearman correlation coefficient [r_s] = −0.7411, P < 0.0001), with modest associations between the presence of validated P. falciparum Kelch13 ART-R mutations (C469Y or A675V) and RSA (median survival 2.6% for wild type [WT] vs 4.1% for mutant, P = 0.01), or eRRSA (median fold change 63.4 for WT vs 30.9 for mutant, P = 0.003) results. Significant correlations were also observed between DHA IC₅₀ values and both RSA percent survival (r_s = 0.4235, P < 0.0001) and eRRSA fold changes (r_s = −0.4116, P < 0.0001). The eRRSA is a scalable alternative for phenotyping fresh P. falciparum isolates, providing similar results with improved throughput.

INTRODUCTION

Artemisinin partial resistance (ART-R) of Plasmodium falciparum, which was first reported in Southeast Asia (1, 2), has recently emerged in eastern Africa (3–7) and threatens the efficacy of artemisinin-based combination therapies for falciparum malaria (8). In Uganda, parasites with five Plasmodium falciparum Kelch13 (PfK13) polymorphisms associated with ART-R, P441L, C469F, C469Y, R561H, and A675V, have emerged (5, 6, 9–11). The ART-R phenotype can be assessed clinically by measuring the parasite clearance half-life (PC_1/2) in patients following treatment with an artemisinin derivative, with a PC_1/2 of >5.5 h generally indicative of ART-R (12, 13). A laboratory marker of ART-R is the ring-stage survival assay (RSA), a microscopy-based measure of the survival of cultured parasites, compared to controls, following a pulse exposure to a high concentration of dihydroartemisinin (DHA) (14, 15).

The RSA was developed for in vitro analysis of laboratory strains and for ex vivo studies of fresh clinical isolates (15, 16). In both cases, the RSA is performed by exposing ring-stage parasites to 700-nM DHA for 6 h, after which cultures are washed free of drug and grown for an additional 66 h. Parasitemias of DHA-treated and control cultures are then enumerated microscopically using Giemsa-stained thin smears (14, 16). This assay is laborious and time consuming, making it challenging to perform at scale (17). In addition, reader subjectivity in distinguishing viable developmentally arrested parasites and dead (pyknotic form) parasites in blood smears leads to assay variability (17, 18). An alternative approach, the extended recovery ring-stage survival assay (eRRSA), was developed to improve quantification and throughput (17). The eRRSA includes the same 6-h exposure of cultures to DHA as in the RSA, but the drug-free incubation after DHA exposure is extended to 114 h to allow parasites to complete an additional intraerythrocytic cycle, improving the differentiation of drug-sensitive (dead) and drug-resistant (surviving) parasites (17). The eRRSA readout utilizes quantitative PCR (qPCR) to quantify the concentration of P. falciparum DNA in treated and untreated control cultures; the extended incubation period was calibrated using known PC_1/2 for a panel of recent patient isolates. This system reduces the workload, allows parallel assessment of multiple samples, can be performed on stored batches of samples, and eliminates the user variability of the RSA, facilitating greater and more precise data acquisition for tracking the emergence and spread of ART-R.

In prior studies of culture-adapted parasite clones from Southeast Asia, the eRRSA provided a stronger correlation with PC_1/2 values (Spearman correlation coefficient [r_s] = −0.8393, P < 0.0001) (17) than did the RSA (r_s = 0.5476) (15). However, to our knowledge, the eRRSA has not been used previously to assess freshly collected clinical isolates. In this study, we compared the RSA and eRRSA for the evaluation of artemisinin susceptibility of fresh P. falciparum isolates collected from study sites in northern and eastern Uganda, where ART-R parasites have emerged.

RESULTS

Baseline characteristics

To compare the artemisinin susceptibility assessment methods in a clinical setting, we collected 122 isolates from patients presenting with uncomplicated P. falciparum malaria at study sites in northern (Agago District, n = 96) and eastern (Tororo and Mbale Districts, n = 26) Uganda. The median (Q1–Q3) age of the participants was 11 years (5–15), and the median (Q1–Q3) sample parasitemia was 1.1% (0.5%–2.5%, Table 1). On average, assays were initiated 23 h after sample collection.

TABLE 1.

Baseline characteristics of participants and assay details^a

Collection period	Number of samples, N		Gender, n (%)		Age (years), median (range)	Sample parasitemia (%)		Time to assay; h
	North	East	M	F		Median (range)	Mean ± SD	Median (range)	Mean ± SD
March–July 2022	96	26	49 (40.2)	73 (59.8)	11.0 (0.5–55)	1.1 (0.2–30.8)	2.1 ± 3.4	23.5 (17.7–27.7)	23.4 ± 1.6

^a F, female; M, male; SD, standard deviation.

Ex vivo DHA susceptibility by RSA, eRRSA, and growth inhibition assays

Of 122 isolates, 117 were assessed by RSA, eRRSA, and standard growth inhibition (IC₅₀) assays; five isolates did not meet our criteria for adequate control growth. With IC₅₀ assays, DHA susceptibility data were available for 106 of the 117 isolates (Table 2; Fig. 1); dose-response curves were uninterpretable for the other 11 isolates. There was a strong negative correlation between RSA survival and eRRSA fold change (r_s = −0.7411, P < 0.0001) (Fig. 2). The nine isolates with RSA percentage survival of >15% (median 22.2%) all had relatively low eRRSA fold changes (median 7.5); five of these isolates, including the three with the highest RSA survival, harbored the PfK13 C469Y or A675V mutation (Table S1). Both RSA survival and eRRSA fold change results had moderate but significant correlations with DHA IC₅₀ values (RSA vs DHA IC₅₀ values, r_s = 0.4235, P < 0.0001; eRRSA vs DHA IC₅₀ values, r_s = −0.4116, P < 0.0001) (Fig. 3).

TABLE 2.

Ex vivo DHA susceptibilities by RSA, eRRSA, and growth inhibition assays^a

Outcome	Mean ± SD	Median (range)
Starting parasitemia (%)	0.8 ± 0.3	1.0 (0.2–1.0)
RSA/eRRSA 72-h control parasitemia (%)	1.1 ± 0.6	1.0 (0.2–4.1)
RSA survival (%)	5.4 ± 7.9	3.0 (0.0–50.0)
eRRSA 120-h fold change	96.8 ± 141.0	46.2 (5.3–900.0)
DHA IC50 (nM)	3.7 ± 1.9	3.2 (1.2–12.3)

^a SD, standard deviation.

Fig 1.

Ex vivo DHA susceptibilities by RSA, eRRSA, and growth inhibition assays. Horizontal lines show medians, and whiskers depict interquartile ranges.

Fig 2.

Spearman rank correlation between RSA and eRRSA results. Dots depict results for individual isolates. WT, wild type.

Fig 3.

Spearman rank correlation between DHA half-maximal inhibitory concentrations (IC₅₀ values) and RSA survival rates and eRRSA fold changes. Dots depict results for individual isolates.

Associations between PfK13 genotypes and drug susceptibility phenotypes

Of the 117 isolates assayed by RSA and eRRSA, we successfully sequenced the pfk13 gene in 103 isolates and characterized PfK13 propeller domain mutations. As expected based on prior findings (5), prevalences of the ART-R-mediating C469Y (14.6%) and A675V (12.6%) mutations were quite high in both northern and eastern Uganda (Table 3). The prevalence of the C469Y mutation was higher in northern compared to eastern Uganda (northern 18% vs eastern 4%, P = 0.004), while that of the A675V mutation was higher in eastern compared to northern Uganda (northern 9% vs eastern 24%, P = 0.019). Two other PfK13 mutations were seen, P441L, which is a candidate marker of ART-R, in one isolate from northern Uganda, and C447R, which has unknown functional significance, in two isolates, one from each region.

TABLE 3.

Prevalence of PfK13 genetic polymorphisms in study isolates by region

PfK13 genotype	Prevalence, n (%)
	Northern (n = 78)	Eastern (n = 25)	P value
PfK13 wild type	55 (70.5)	17 (68.0)	–
PfK13 polymorphism
C469Y	14 (18.0)	1 (4.0)	0.004
A675V	7 (9.0)	6 (24.0)	0.019

Considering the RSA, eRRSA, and growth inhibition assays, PfK13 C469Y and A675V mutant parasites exhibited lower susceptibility than wild type (WT) parasites, although not all differences were significant (WT vs C469Y: RSA median survival 2.6% vs 3.0%, P = 0.306; eRRSA median fold change 63.4 vs 32.1, P = 0.049; DHA IC₅₀ 3.0 nM vs 3.9 nM, P = 0.026; WT vs A675V: RSA median survival 2.6% vs 6.9%, P = 0.005; eRRSA median fold change 63.4 vs 27.0, P = 0.008; DHA IC₅₀ 3.0 nM vs 5.1 nM, P = 0.098) (Fig. 4).

Fig 4.

Associations between PfK13 genotypes and drug susceptibility phenotypes. RSA survival, eRRSA fold changes, and DHA IC₅₀ values for wild type (WT) and mutant parasites. The horizontal lines represent medians, and error bars represent interquartile ranges. Kruskal-Wallis P values indicate that significant variation was observed among the samples for each phenotype; differences between each pair of samples within each group were assessed using Mann-Whitney, with significant differences (P < 0.05) shown in bold.

DISCUSSION

The emergence of ART-R in eastern Africa necessitates robust, fast, and reliable measurement of drug resistance phenotypes in field settings. The RSA (14), the standard laboratory measure of DHA susceptibility phenotypes, has been critically important, but it has significant challenges, in particular low throughput and user variability in microscopy readouts (17, 18). The eRRSA was developed to improve throughput and reduce variability, allowing for rapid and reliable processing of many samples (17). In this study, the first evaluation of the eRRSA using fresh clinical isolates, results from the eRRSA and RSA were highly correlated across a wide range of parasitemias. Correlations were also seen with results from standard DHA growth inhibition assays and with genotypes associated with ART-R. Thus, the eRRSA offers a practical, scalable alternative to the RSA for phenotyping fresh P. falciparum isolates.

The eRRSA was designed to offer a readout (qPCR) that greatly decreases user workload and the potential for variability in readings compared to the RSA, which relies on microscopic expertise to count and distinguish viable and non-viable parasites. After culture for an additional life cycle before eRRSA readouts, contributions of non-viable parasites are presumed to be minimal, with qPCR results reflecting viable parasites. As seen with laboratory isolates (17), we found excellent correlation between eRRSA and RSA results. These assays showed modest correlation with DHA IC₅₀ results. Although the DHA IC₅₀ assay is not considered a reliable marker for ART-R (2, 15, 19), it appears that DHA IC₅₀ values are increased in ART-R parasites in Africa (11). The eRRSA and RSA results also generally correlated with PfK13 genotypes, with both of the validated ART-R mutations now prevalent in northern and eastern Uganda, C469Y and A675V, associated with decreased susceptibility. These results all suggest that the eRRSA offers an acceptable and practical replacement for the RSA.

A valuable aspect of the eRRSA is its scalability. Challenges with scale have limited the use of the RSA, as each assay requires extensive reader time. The eRRSA is much simpler but requires access to qPCR, which is now a basic method in most molecular laboratories. Importantly, for the eRRSA, qPCR need not be performed at a clinical site but can be performed at a central facility using DNA that is stored or extracted from dried blood spots. The scalability of the eRRSA will be beneficial as work to better understand the biology of ART-R in Africa continues. This simpler assay can readily allow increased assay points and exploration of more assay parameters (e.g., duration of incubation and concentration of DHA) to better characterize ART-R.

The eRRSA is not the only modification in the RSA to offer simplified methodology. For the standard RSA, counting of parasites using both a DNA-binding dye (SYBR green I) and a mitochondrial dye (MitoTracker deep red) allowed automated counting of viable parasites by flow cytometry, obviating the need for microscopy counts (18). Another study utilized two DNA-binding fluorophores (ViSafe Green and propidium iodide) and a fluorophore-conjugated human leukocyte antibody (allophycocyan-tagged anti-CD45) to identify viable parasites (20). However, both of these approaches require either access to flow cytometry on site or cell preservation for subsequent off-site assays.

This study identified some interesting drug susceptibility results, including isolates with surprisingly low susceptibility to DHA based on either growth inhibition or RSA assays and relatively poor correlations between the presence of validated ART-R PfK13 mutations and results with either RSA. These results are consistent with other recent ex vivo data from Uganda (11, 21) and might be explained by the complexities of studying polyclonal field isolates and differences in the molecular mediators of drug susceptibility between Asian and African malaria parasites, including potential impacts of polymorphisms in addition to PfK13 mutations on DHA susceptibility.

Our study had some limitations. First, a relatively small number of samples was available for assays. However, the numbers were adequate to identify a strong association between eRRSA and RSA results. Second, measures of clinical parasite clearance were not available for these isolates, preventing us from directly validating the eRRSA as an ex vivo marker of clinical ART-R. Third, although we studied isolates from two distant sites, these were from only one region of Africa, and results might not be fully representative of malaria parasites around the world.

IIn summary, in freshly isolated malaria parasites, the newly developed eRRSA and the RSA, the standard laboratory assay to identify ART-R in P. falciparum, yielded similar results. These phenotypic laboratory markers of ART-R were associated with previously validated genetic markers of ART-R, the C469Y and A675V PfK13 mutations that are now circulating in Uganda. The eRRSA offers a simpler, higher-throughput assay compared to the RSA, suggesting that it can replace the RSA in field settings where laborious slide reads or flow cytometry readouts are impractical, but where qPCR is available on site or at an accessible reference laboratory.

MATERIALS AND METHODS

Collection and processing of P. falciparum Isolates

Blood samples were collected from 122 symptomatic patients with uncomplicated malaria at Patongo Health Center III in Agago District in northern Uganda or Tororo District Hospital in Tororo District and Busiu Health Center IV in Mbale District, both in eastern Uganda, from 24 March to 31 July 2022. Enrollment criteria included age 6 months or older, P. falciparum monoinfection confirmed by a Giemsa-stained blood film, no signs of severe disease, and written consent (with parental consent for children <18 years and assent for children 8–17 years).

To initiate drug susceptibility assays, 2- to 5-mL blood samples were collected into heparin tubes by venipuncture and transported on ice on the day of collection to a laboratory in Tororo, where they were stored at 4°C overnight and assayed the following day, generally within 24 h of sample collection. For the assays, samples were centrifuged at 2,000 rpm for 10 min, and plasma and buffy coats were removed. Erythrocyte pellets were washed 3× in Roswell Park Memorial Institute (RPMI) wash media (RPMI 1640 with L-glutamine and 25-mM HEPES [Thermo Fisher Scientific], 0.2% NaHCO₃, and 10 mg/L gentamicin [Gibco]). Pellets were reconstituted to 50% hematocrit using complete culture media (wash media plus 0.1-mM hypoxanthine and 0.5% AlbuMAX II [Thermo Fisher Scientific]). One hundred microliters of washed pellet was also spotted on Whatman filter paper (Cytivia) for subsequent study of DNA. Parasitemia was determined by counting the proportion of parasite-infected erythrocytes per 1,000 erythrocytes on Giemsa-stained thin blood films. Samples with morphological evidence of non-falciparum species were excluded.

RSA

The ex vivo RSA was performed as previously described (16). Briefly, parasite samples with parasitemia of 0.2%–1.0% were cultured at their original parasitemia and 2% hematocrit; those with parasitemia of >1% were diluted using complete culture media and donor erythrocytes to 1% parasitemia and 2% hematocrit. Two milliliters of culture was aliquoted into 2 wells of a flat-bottomed 24-well culture plate. One well was incubated with 700-nM DHA (supplied by Medicines for Malaria Venture) dissolved in dimethyl sulfoxide (DMSO), and the other with 0.1% DMSO (the same concentration as in the treated samples) for 6 h in 90% N₂, 5% CO₂, and 5% O₂ at 37°C in a modular incubator (Billups Rothenberg), followed by washing 3× with 10 mL of prewarmed RPMI wash media. The cultures were then resuspended in 2 mL of RPMI complete culture media and incubated for an additional 66 h. At 72 h (66 h post drug removal), old culture media were aspirated and replaced with an equal volume of fresh RPMI complete culture media. At this point, the assays diverged (Fig. S1). For the RSA, 300 μL was used for preparation of Giemsa-stained thin blood films, and parasitemia was determined by counting a minimum of 2,000 erythrocytes for DMSO-treated controls and 20,000 erythrocytes for DHA-treated cultures. RSA survival was expressed as the proportion of viable parasites in the DHA-treated cultures relative to DMSO controls. Higher levels of resistance are associated with increasing percentage survival.

ERRSA

The eRRSA was continued after the RSA sampling and followed our previously described protocol (17). The remaining cultures were incubated as described above for an additional 48 h (total 120 h), after which 8 µL was collected from control and DHA-treated cultures and spotted onto filter papers for qPCR amplification. To ensure that only isolates with good ex vivo growth were assayed, RSA and eRRSA results were considered valid only for samples that had ≥0.2% initial parasitemia, ≥0.2% 72-h control parasitemia, and 72-h control parasitemia ≥25% of the initial parasitemia. Parasite DNA was eluted by incubating filter paper dried blood spots in Phusion Blood Direct 2× PCR buffer (Thermo Fisher Scientific) for 30 min with shaking at 60°C. qPCR was performed as previously described for liquid samples (17) but modified for dried blood spots by substituting 5 µL of the Phusion Blood Direct 2× buffer/DNA mix as the template, supplemented with 1×  SYBR green I in a 10-µL reaction. DNA was amplified using optimized forward and reverse primers encoding the pfcrt gene. qPCR amplification was measured using the fast mode of the ABI 7900HT, with a 20-s denaturation at 95°C, followed by 40 cycles of 95°C for 1 s, 62.3°C for 30 s, and 65°C for 15 s. eRRSA values were expressed as fold change, reflecting the difference in DNA content between treated and untreated cultures. Cycle threshold (Ct) values were calculated using the ABI SDS 2.4.1 and converted to fold change (2^ΔCt) by determining the average ΔCt for three technical replicates for the untreated and treated samples by applying the following equation:

$${\displaystyle \mathrm{F}\mathrm{o}\mathrm{l}\mathrm{d}\mathrm{c}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{g}\mathrm{e}=2^{(\mathrm{a}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{g}\mathrm{e}\mathrm{C}\mathrm{t}\mathrm{o}\mathrm{f}\mathrm{D}\mathrm{H}\mathrm{A}\mathrm{t}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}-\mathrm{a}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{g}\mathrm{e}\mathrm{C}\mathrm{t}\mathrm{o}\mathrm{f}\mathrm{D}\mathrm{M}\mathrm{S}\mathrm{O}\mathrm{t}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}).}}$$

Higher levels of resistance are associated with decreasing fold change (differences in Ct between treated and untreated samples).

Ex vivo growth inhibition assays

DHA IC₅₀ values were determined using a 72-h 96-well microplate growth inhibition assay with drug concentrations optimized to capture full dose-response curves, and with SYBR green detection, as previously described (11, 21) (Fig. S2). Cultures were diluted with uninfected erythrocytes from local blood banks for total volumes of 200 µL per well at 0.2% parasitemia and 2% hematocrit. Culture plates were incubated under the conditions described above for RSAs. After 72 h, well contents were resuspended, and 100-µL culture per well was transferred to black 96-well plates containing 100 µL per well SYBR Green lysis buffer (20-mM Tris, 5-mM EDTA, 0.008% saponin, 0.08% Triton X-100, and 0.2-µL/mL SYBR Green I [Invitrogen, Thermo Fisher Scientific]) and mixed. Plates were incubated for 1 h in the dark at room temperature, and fluorescence was measured with a FLUOstar Omega plate reader (BMG LabTech, 485-nm excitation and 530-nm emission). IC₅₀ values were derived by plotting fluorescence intensity against log drug concentration and fit to non-linear curves using a four-parameter Hill equation in Prism (version 9.3), as previously described (21). IC₅₀ values from cultures with poor growth (fluorescence range <3,000 fluorescence index) were excluded from analysis.

PfK13 genotyping

Parasite DNA was extracted using Chelex-100 and 0.01% Tween-20 as previously described (6, 11). Pfk13 gene sequencing was performed with molecular inversion probe (MIP) capture, library preparation, and sequencing, as previously described (21, 22). Briefly, template DNA and a panel of drug resistance probes were incubated with DNA polymerase and ligase enzymes to enable the MIP capture reaction. Uncircularized probes and template DNA were removed using an exonuclease, leaving only captured circularized DNA in the wells. Adapters and sample barcoded DNA primers were added, and target DNA was PCR amplified. Gel electrophoresis of barcoded DNA was performed to remove unbound adapters and primer dimers before pooling into a sequencing library. Sequencing was performed on the illumina MiSeq platform.

Statistical analysis

The relationship between RSA survival rates and eRRSA qPCR fold changes was evaluated using Spearman’s rank correlation, and RSA survival percentage, eRRSA fold changes, and DHA IC₅₀ values of PfK13 wild-type and mutant parasites were compared using the Kruskal-Wallis and Mann-Whitney U tests, with significance at P < 0.05. Prevalences of PfK13 genetic markers of ART-R were compared between regions using Fisher’s exact test. Statistics were performed and figures were generated using GraphPad Prism (version 10.2.3).

ETHICS APPROVAL

The study was approved by the Makerere University School of Biomedical Sciences Research and Ethics Committee, the Uganda National Council for Science and Technology, and the University of California, San Francisco Committee on Human Research.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aac.01183-24.

Supplemental material. aac.01183-24-s0001.docx.

Table S1; Figures S1 and S2.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.