Background

Intensive malaria control efforts have yielded progress toward malaria elimination in multiple endemic countries [1]. In some settings, as the burden of Plasmodium falciparum and Plasmodium vivax malaria declines, less common species such as Plasmodium malariae and Plasmodium ovale have increased in prevalence as well as public health relevance [2,3,4]. Malaria elimination requires rapid detection and treatment of all Plasmodium species. However, existing rapid methods are not species-specific and have poor sensitivity for these less common species [5]. Though P. malariae is known to cause chronic parasitaemia, sometimes lasting years, and causing chronic anaemia and splenomegaly [6], its treatment is straightforward. Thus, its detection and treatment can reduce chronic carriage and contribute to malaria elimination efforts. This study reports a new P. malariae-specific diagnostic assay that uses rapid, isothermal recombinase polymerase amplification (RPA) and lateral flow strip detection and has potential for further development into a point-of-care tool.

Methods

Primer and probe design

Primers and probes were designed according to manufacturer-suggested best practices [TwistDX(TwistAmpTM nfo), Cambridge, UK]. Primers were designed to meet the following parameters: 30–36 nucleotide (nt) length, 40–50% GC content, 50–100 °C melting temperature, < 5nt mononucleotide repeat length, and 80-500nt amplicon length. Publicly available Pm 18S ribosomal RNA gene sequences from PlasmoDB and NCBI databases were used to design primers using Primer3Plus v3.2. Primers from published literature were also modified and used to guide selection of target regions [7,8,9]. Probes were designed within the target sequence, with 46-52nt length, 20–80% GC content, and 57–80 °C melting temperature.

RPA assay protocol

RPA reactions were performed using a reverse primer biotinylated on the 5′ end and an unlabelled forward primer; a 5′-FAM-labelled probe with an abasic residue and 3′ blocker modification. The blocker prevents extension until cleavage of the abasic site by Endonuclease IV (nfo) enzyme. The biotinylated primer and FAM labelled probe form a duplex of double-stranded RPA amplicons that are detectable by sandwich assay, involving a lateral flow strip test band containing anti-FAM antibodies. Direct visualization by naked eye is possible using streptavidin-gold conjugates that bind biotinylated double stranded amplicon captured by anti-FAM antibodies on the lateral flow strip (Fig. 1).

Fig. 1
figure 1

Plasmodium malariae lateral-flow based RPA assay schematic. (1) Reaction components include unlabelled forward and 5′-biotinylated reverse primers, 5′-FAM labelled probe with an abasic residue and 3′ blocker, endonuclease IV enzyme (nfo), and template DNA. (2) RPA reaction proceeds via primer-recombinase-SSB complex de-looping, nfo cleavage, polymerase extension that results in FAM- and biotin-labelled double-stranded amplicons. Exponential amplification occurs during 20-min isothermal incubation at 39 °C. (3) Lateral flow strips with a band containing anti-FAM antibodies are loaded with RPA product diluted in buffer containing streptavidin-conjugated gold nanoparticles. (4) Detection of labelled RPA product immobilized on the lateral flow strip is performed by visual inspection using the naked eye. Total reaction time is approximately 35 min. Figure made using BioRender

Full size image

The RPA assay was performed as per manufacturer instructions with slight modifications and using 10 µL reaction volumes (comprising 9 µL of master mix and 1 µL of DNA template). Forty five µL of master mix per five reactions (or proportional equivalent) was prepared, containing 29.5 μL RPA-nfo rehydration buffer, 2.5 μL 280 nM MgOAc, 2.1 μL forward primer (10 μM) and 2.1 μL reverse primer (10 μM), 0.6 μL of probe (10 μM), 1 μL (200 ng/μL) human DNA (for non-clinical samples), and 6.2–7.2 µL molecular-grade water. Reactions were incubated at 39 °C for 30 min with brief intermittent vortexing after 4 min. After amplification, 2 μL of each reaction was immediately diluted with 98 μL of wash buffer in an Eppendorf tube, and a disposable lateral flow strip (Ustar Biotechnologies Ltd, Hangzhou, China REF: D001-03) was dipped into the tube. The result was recorded after 5 min, although results are usually apparent earlier. A test was considered positive when both the test line and control line visualized by the naked eye; and negative only when the control line is visible. When no line appeared, the result was considered invalid. Samples with invalid results were retested.

Assay optimization

Both TwistDX basic and TwistDX nfo kits were used during optimization of reaction conditions before evaluation of assay performance. TwistDX basic products were denatured using heat (60 °C for 10 min) or purified using commercially available DNA purification kit (Qiagen, Germantown, USA) and visualized by 3% agarose gel electrophoresis. The RPA reaction was performed according to manufacturer’s protocol except with temperature ranging from 34–40 °C for 30 min. For non-clinical samples, 18S rRNA plasmid DNA was combined with 200 ng of human DNA to simulate human sample collection.

Assessment of assay performance

To determine analytical sensitivity and specificity of the assay, 18S rRNA plasmid DNA of P. malariae (MRA-179, Lot: 70031271), P. falciparum (MRA-152-G, Lot: 59201399), P. vivax (MRA-178, Lot: 58067149), and P. ovale (MRA-180, Lot: 70043212) obtained from MR4 (BEI Resources, Manassas, USA) was used. To determine limit of detection, P. malariae 18S rRNA plasmid DNA was diluted from 106 to 10–1 copies/μL. Analytical specificity was evaluated using 18S rRNA DNA from other Plasmodium species at high concentration (P. falciparum and P. vivax at 106 copies/µL, P. ovale at 103 copies/µL). Each reaction (RPA and lateral flow detection) was performed with a minimum of four replicates.

Clinical sensitivity and specificity were evaluated using genomic DNA from patients infected with P. malariae, P. falciparum, P. ovale spp., and/or P. vivax collected as part of studies in Tanzania, Democratic Republic of Congo (DRC), and Ethiopia [10,11,12]. Clinical samples were tested using RPA performed in singleton; all sensitivity and specificity estimates were determined using these results. Posthoc, repeat testing was performed for one P. vivax sample. For all studies, participants provided informed consent or parental consent was obtained at the time of enrollment. DNA was extracted from dried blood spot samples using Chelex-100, and evaluated for P. malariae, P. falciparum, P. ovale spp., and P. vivax infection using real-time PCR targeting the 18S rRNA gene as previously described [10].

Results

Selection of primers, probe, and optimized conditions

Fifteen primer pairs were evaluated for the current P. malariae RPA assay (listed in Additional file 1: Table S1). The length of the primers tested ranged between 30 and 41 nt, with melting temperature ranging from 51 to 62.4 °C and GC content ranging from 20 to 41%. The best performing primers and probe during initial testing using TwistDX basic kits were selected for further evaluation and validation using the TwistDx nfo kit: primers AJMP_7: 5′-ATAACATAGTTGTACGTTAAGAATAACCGC-3′ (forward), AJMP_30: 5′- ATATATAATACTTCGATTAGTTGAGTACCT-3′ (reverse), and probe AJMP_42: 5′- GTTGTACGTTAAGAATAACCGCCAAGGCTTTATTTTTTCTGTTAC-3′. Primer and probe modifications are described in Additional file 2: Table S2. The optimized RPA assay reaction consisted of two-steps (Fig. 2): (1) RPA assay in a heat block at 39 °C for 30 min, and (2) lateral flow assay detection with visualization of results after 5 min (35–45-min total reaction time, not including DNA extraction or reagent preparation).

Fig. 2
figure 2

Assay workflow and timeline. Sample collection, DNA extraction (and thawing of DNA and reagents, if necessary) times vary by method. Newer methods enable rapid extraction in ≤ 10 min (e.g., QuickExtract DNA, Biosearch Technologies, Hoddesdon, UK), but at higher cost, whereas conventional Chelex extraction can require up to 24 h at lower cost. Incubation temperature can be varied, though 39 °C was optimal for the assay. Visualization of positive results is often possible in < 2 min but reported at 5 min to standardize assay output. Figure made using BioRender

Full size image

Analytical sensitivity and specificity

The assay’s analytical sensitivity approached the sensitivity of P. malariae-specific PCR assays [8] and achieved perfect analytical specificity (Fig. 3A, B). When tested against serial dilutions of P. malariae 18S rRNA plasmid DNA, the lower limit of detection was between 10 and 100 copies/µL (~ 1.7–17 genome equivalents/µL, assuming six copies of 18S rRNA per genome [12, 13], with all replicates containing ≥ 10 P. malariae copies/µL detected except for a single replicate. Specificity assessed using high-concentration plasmid DNA for P. falciparum (106 genomes/µL), P. vivax (106), and P. ovale spp. (103) confirmed 100% analytical specificity.

Fig. 3
figure 3

Sensitivity and specificity of the P. malariae RPA-lateral flow (RPA-LF) assay. A Analytical sensitivity and specificity determined using serially diluted 18S rRNA plasmid copies/µL (equivalent to approximately 1700–0.17 genome equivalents/µL). B Example lateral flow read-out by species (Pf, P. falciparum; Pv, P. vivax; Po, P. ovale spp.; Pm, P. malariae; NTC, no-template control). C Clinical sensitivity determined using 21 Tanzania field samples with qPCR-confirmed P. malariae infection with a range of qPCR-determined parasite densities (log10). One low-density case was missed (*). Three cases of Pm + Pf co-infection were included (^); the parasite density indicated for these Pm + Pf cases refers only to Pm. D Clinical specificity determined using Tanzania and Ethiopia field samples with qPCR-confirmed P. falciparum, P. ovale, and P. vivax

Full size image

Clinical sensitivity and specificity

Clinical sensitivity and specificity assessed using field samples from Tanzania and Ethiopia were excellent, with only a single false-negative and false-positive, respectively (Fig. 3C, D). The assay’s clinical sensitivity compared to qPCR was 95.2%, detecting 20 of 21 P. malariae qPCR-confirmed field samples (17 mono-infected and three mixed P. falciparum and P. malariae co-infection) from Tanzania with a range of parasite densities down to 100 and 10–1 parasites/µL, corresponding to Ct values ranging 22–41. A single false-negative result was observed in P. malariae field sample with parasite density of 0.5 parasites/µL, below the assay’s analytical limits of detection. Clinical specificity was 88% during testing against three P. falciparum, two P. ovale spp., and three P. vivax field samples from Tanzania and Ethiopia. One false-positive result was observed in a P. vivax field sample from Ethiopia, with a faint positive band that was not visualized during repeat posthoc testing.

Discussion

This study describes a new P. malariae RPA-lateral flow assay with strong performance when applied to laboratory and field samples, and potential for development for future use in the field. Species-specific point-of-care assays for neglected Plasmodium species such as P. malariae are not currently available. PCR assays for these species require thermocyclers and materials that are not available in many settings where malaria is endemic. This P. malariae RPA-LF assay provides a simple, sensitive, and specific option to detect P. malariae in laboratories in low-resource settings.

The P. malariae-specific RPA-LF assay achieved limits of detection 10–100 fold better than existing pan-Plasmodium lactate dehydrogenase rapid diagnostic tests, with a lower limit of detection of 1.7–16.7 versus 100–1000 parasites/µL, respectively [14]. The assay had perfect analytical specificity and yielded only a single false-negative during testing against field samples. The false-negative result was observed in a sample with low parasite density below the assay’s limit of detection. While the assay detected several other field samples with lower parasite densities, its performance against low-density samples ≤ 10 parasites/µL is expected to be less robust than higher-density samples, but this is also the case with qPCR. Assay performance was comparable to other laboratory-based molecular methods such as PCR and loop-mediated isothermal amplification (LAMP) for P. falciparum (Table 1). It was also less complex than CRISPR-based assays that incorporate RPA, though with lower sensitivity for P. malariae detection [15, 16]. The assay had excellent specificity, achieving perfect results versus high-concentration plasmid DNA. Its single false-positive during testing of field samples involved a faintly positive test band that was not observed during repeat testing performed as part of post hoc analysis. This suggests that the initial false-positive call was likely a result of contamination during sample processing and RPA-lateral flow testing.

Table 1 Comparison of malaria diagnostic methods (adapted from Mbanefo and Kumar 2020)
Full size table

RPA provides a rapid method of amplification, reportedly faster than other isothermal amplification methods [17], that can be performed at/near human physiological temperatures. The assay had a total reaction time of less than 40 min and performed best at 39 °C, suggesting opportunities for future use without a heat block (for example, incubation in the axilla or near a simple stove), though these approaches are not yet tested.

Several limitations must be overcome before the current assay is ready for point-of-care use. First, reagent stock outs hampered RPA-lateral flow diagnostic development throughout the COVID-19 pandemic and have not been rectified. Alternatives to TwistDx’s proprietary approaches and kits are now becoming available [18]. Second, the assay requires DNA input and cannot be applied directly to whole-blood samples in its current form. DNA extraction is required prior to performance of the RPA-lateral flow assay and adds time and complexity to sample processing. Advances in simple DNA extraction methods promise rapid turnaround time (< 10 min), but nonetheless add complexity to the workflow. One-pot reactions that eliminate this requirement would simplify assay requirements and improve assay performance. Third, though the assay outperforms pan-Plasmodium RDTs and was on par with conventional real-time PCR, it could still miss low-density P. malariae infections. This is important because P. malariae infections have lower parasite densities than P. falciparum and can be missed by commonly used PCR assays. However, it is still likely to pick up the majority of P. malariae parasitemia’s detectable by qPCR as well as those contributing to febrile illness [4]. The current assay outperforms pan-Plasmodium RDTs, and a future point-of-care version could fill a key gap in the malaria diagnostic portfolio.

Evidence is limited about P. malariae, including its true prevalence and distribution, and the extent of its propensity to cause chronic carriage and clinical disease. Simpler diagnostic tools that are closer to the point of care could help resolve some of these questions. They may also contribute to efforts to achieve malaria elimination, as all malaria species need to be detected and addressed. The P. malariae-specific RPA-LF assay described here provides a useful laboratory tool for countries seeking to address a key neglected malaria species known for chronic parasitism in low resource settings.