Evaluation of a novel real-time PCR assay for the detection, identification and quantification of Plasmodium species causing malaria in humans

The analytical specificity of the assay was determined by testing samples received from reference institutes or referral laboratories [the Dutch Foundation for Quality Assessment in Medical Laboratories (SKML), UK NEQAS, Streeklab Haarlem—Laboratory for Medical Microbiology, the Biomedical Primate Research Centre, the London School of Hygiene & Tropical Medicine, and Labor Dr. Wisplinghoff]. In total, 40 reference samples were tested that included all seven Plasmodium species (P. falciparum, P. vivax, P. knowlesi, P. malariae, P. cynomolgi, P. ovale wallikeri, and P. ovale curtisi) at various levels of parasitaemia, and four different non-Plasmodium parasites: Babesia microti, Loa loa, Trypanosoma brucei rhodesiensis, and Leishmania donovani infantum. Specimen types included ethylenediaminetetraacetic acid (EDTA) whole blood, isolated DNA, and Giemsa-stained blood or bone marrow slides (Table 1). The results of the MC004 assay were compared with the results of the reference laboratory.

The MC004 assay (MRC-Holland, Amsterdam, the Netherlands) is a multiplex nucleic acid amplification test that detects mitochondrial Plasmodium DNA encompassing the cyclo-oxygenase 3 (COX-3), cyclo-oxygenase 1 (COX-1) and cytochrome b (CYTB) genes from Plasmodium species that cause malaria in humans. The assay provides a qualitative result for the presence/absence of Plasmodium DNA, and discriminates between 11 Plasmodium species/strains (P. falciparum, P. vivax, P. malariae, P. ovale wallikeri, P. ovale curtisi, P. knowlesi LT48, P. knowlesi ATCC 30153, P. knowlesi ATCC 30158 P. cynomolgi ATCC 30149, P. cynomolgi KJ569866.1 and P. cynomolgi KJ569868.1). The MC004 assay involves two main steps, which take place in a single tube: (1) asymmetric target amplification by two different primer sets (primer pair 1: 5ʹ-TCGCTTCTAACGGTGAACT-3ʹ/5ʹ-AAGCAAACACTAGCGGTGGAA-3ʹ and primer pair 2: 5ʹ-CAGTATAATATTGTAATTTGATCAGTATGAG-3ʹ/5ʹ-GGATATTGTATAAATGATGCTATATCAGGTA-3ʹ). Primer pair 1 was designed to amplify a 116 base pair fragment of all 11 Plasmodium species/strains. Primer pair 2 was designed to amplify a 212 base pair fragment of P. vivax, P. knowlesi, and P. cynomolgi only. See Table 2 for the precise annealing positions of the primers. Amplification can be monitored real-time in the FAM detection channel due to the presence of a nonspecific green fluorescent nucleic acid dye that becomes fluorescent upon binding to double stranded DNA. (2) detection and differentiation of the amplicons by molecular beacon probe-based melting curve analysis. The assay uses three different probes, which are either Texas red-labelled (identical to the sequence of P. knowlesi ATCC 30153), Cy5-labelled (identical to the sequence of P. falciparum), or Cy5.5 labelled (identical to the sequence of P. knowlesi LT48).

The preferred specimen type is DNA extracted from 200 µL of (EDTA) human whole blood. DNA was isolated using the QIAamp DNA Blood Mini QIAcube Kit (Qiagen, Hilden, Germany) on a QIAcube instrument (Qiagen, Hilden, Germany) following the manufacturer’s specifications. DNA was eluted with 100 µL of elution buffer. Giemsa-stained blood smears were processed by pipetting 10 µL of 10 mM sodium hydroxide (NaOH pellets, 99%, Merck, Darmstadt, Germany) onto the surface of the blood smear. Subsequently, blood smear material was scraped off from the glass slide by making circular movements with a sterile lancet (Solofix, B Braun, Oss, Netherlands). The collected specimen was transferred to a sterile 2.0 mL tube (Eppendorf, Hamburg, Germany) that contained 100 µL of 10 mM sodium hydroxide. Processed specimen samples were stored at − 30 °C.

Each reaction was carried out in a final volume of 25 µL reaction mixture containing 23 µL of MC004 master mix (MRC-Holland, Amsterdam, the Netherlands) and 2 µL of template. The components of the master mix were: both forward primers (1.25 µM), both reverse primers (0.25 µM), all three probes (0.5 µM), fluorescent compound, dNTPs, reaction buffer, and Taq DNA polymerase. The following settings were used for the amplification step: 95 °C for 3 min, followed by 50 cycles of 95 °C for 15 s, 60 °C for 30 s and 68 °C for 40 s. The following settings were used for the melting curve step: a gradual temperature increase from 25 to 69.4 °C (0.4 °C per 5 s).

Positive and negative control samples were used to validate run validity. Each run included two negative controls: (1) with elution buffer added as template and (2) a fresh uninfected whole blood specimen processed as a separate sample. Positive control samples for each Plasmodium species come with the MC004 assay as provided by MRC Holland. The 1st WHO International Standard for P. falciparum DNA (NIBSC code: 04/176) diluted in EDTA whole blood (before DNA isolation) to a concentration of 1 × 10^–3 IU/mL was used as a positive control for limit of detection. Each sample should produce an amplification curve, if no amplification curve is observed, this may be a sign of inhibition and renders the result invalid.

The assay was performed on the CFX96 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). Runs were analysed using the accompanying Bio-Rad CFX manager 3.1 software (Bio-Rad, Hercules, CA, USA). Cq-values were determined by the software’s Cq Determination Mode (settings: Single Threshold set at 2000; Baseline Setting: Baseline Subtracted Curve Fit).

Melting temperatures were determined using the software’s default settings. The throughput is approximately 80 specimens for 1 operator in about 3.5 h. Up to 12 specimens can be processed, including DNA isolation using the QIAcube, within four hours of arrival of the specimens at the laboratory.

The limit of detection of the MC004 assay was determined by preparing a tenfold serial dilution of the WHO International Standard for P. falciparum DNA (NIBSC, Hertfordshire, England) in uninfected EDTA whole blood to achieve a concentration range of 1 × 10⁹ IU/mL to 1 × 10^–9 IU/mL. To clarify, this WHO standard was assigned an arbitrary unitage in International Units (IU), unrelated to the level of parasitaemia or DNA copy number [19]. The lyophilized WHO standard, reconstituted in 500 µL of sterile water, was diluted in EDTA whole blood. Total volume of each dilution was 400 µL. DNA was isolated as described above, and samples were tested in three independent runs. The limit of detection was determined as the lowest concentration of P. falciparum detected in all three runs. In addition, similar tenfold serial dilutions of patient samples with known parasitaemia were tested: P. falciparum 1.3%, P. falciparum 0.2%, P. vivax < 0.1%, P. malariae 0.3%, P. ovale curtisi < 0.1%, and P. knowlesi ATCC 30158 0.3%. Samples were tested in two independent runs. The limit of detection was determined as the lowest concentration of the Plasmodium species detected in both runs.

Mixed Plasmodium infections were created by adding together two DNA samples, each containing a different Plasmodium species (P. falciparum and a non-falciparum species) 1:1 (equal volume). When applicable, the parasitaemia was determined by experienced microscopists (P. falciparum 0.2%, P. vivax 0.1%, P. ovale wallikeri 0.2%, P. malariae 0.4%, P. knowlesi ATCC 30158 0.3%, P. ovale curtisi and P. cynomolgi KJ569868.1 isolated DNA diluted 100 times with sterile water). In addition, a two-fold serial dilution (1:2, 1:4, 1:8, 1:16, 1:32, 1:64, 1:128) of a P. falciparum DNA sample (initial parasitaemia was 1.3%) in sterile water was prepared to create mixed infections of P. falciparum and a non-falciparum species (non-falciparum samples as described above).

To quantify parasite density, a calibration curve was established by plotting the microscopically determined parasite densities of ten samples from patients with P. falciparum infections as a function of the Cq-value. The parasite densities were: 0.1%, 0.1%, 0.4%, 0.8%, 1.3%, 1.5%, 3.7%, 23.7%, 31.8%, and 35.8%. Cq-values were determined in three independent runs. The exponential equation (y = a*e^(b*x)) of the line of best fit through the 30 data points was determined using Microsoft Excel for Mac 2016, where ‘y’ is the parasitaemia (percentage of infected erythrocytes), ‘x’ is the Cq-value (threshold cycle), ‘a’ and ‘b’ are the dimensionless coefficients that define the equation, and ‘e’ is the base of the natural logarithm.

Within-run and between-run precisions [standard deviations (SD) and coefficients of variation (CV)] for the quantification of the parasitaemia were determined using four clinical samples infected with P. falciparum. The parasitaemia levels of these four samples were estimated by real-time PCR to be around 0.1%, 0.2%, 2% and 27%, respectively. Precision was assessed by testing five replicates per run over five runs. Each run the samples were tested in different wells of the PCR plate. Within-run and between-run precision were calculated as described by Chesher, 2008 [20].

Probe melting temperatures for each Plasmodium species/strain are shown in Table 3. The observed maximum deviation of melting points was 1.2 °C. Figure 1 shows representative melting curves of P. falciparum and P. vivax (see Additional file 1: Figure S1 for the melting curves of all other Plasmodium species/strains).

Testing of samples that did not contain any parasites and samples that contained Leishmania donovani infantum, Loa loa, Trypanosoma brucei rhodesiensis, or Babesia microti, yielded no false positive results. Among the samples that did contain Plasmodium parasites, one discrepancy was observed between the result of the reference institute [see Table 1: SKML, P. vivax (2018.1G)] and the MC004 assay (P. knowlesi). This sample was further analysed by Sanger sequencing, and this confirmed the presence of P. knowlesi. No data regarding the origin of this specimen were available at the reference institute, therefore leaving the question of what might explain this anomaly open-ended. The analytical specificity panel contained one P. cynomolgi sample, which showed melting temperatures consistent with P. cynomolgi KJ569868.1. Five of the six P. knowlesi samples showed melting temperatures consistent with P. knowlesi LT48, the other one with P. knowlesi ATCC 30158. The one included mixed infection (UK NEQAS 5844, P. falciparum + P. vivax) was readily identified. Thus, among the 40 samples that were included in the specificity panel, no false positive or false negative results were observed with the MC004 assay.

Limit of detection results are presented in Table 4. The 1st WHO International Standard for P. falciparum DNA (NIBSC code: 04/176) was detectable until a dilution of 10^–12, which corresponds to a concentration of 1 × 10^–3 IU/mL. The undiluted concentration of this standard has been described to correspond to 9.79 parasites per 100 red blood cells [21]. However, arguably, expressing the limit of detection as number of parasites per volume of blood would be misrepresentative, as the limit of detection determined by dilution series appears to be dependent on the starting concentration. Illustrative of this point are the dilutions of P. falciparum with a relatively high parasitaemia of 1.3%, which was still detectable at dilution 10^–14, and of P. falciparum with a relatively low parasitaemia of 0.2%, which was detectable until dilution 10^–5. Limits of detection of P. vivax, P. knowlesi ATCC 30158, and P. ovale curtisi were determined at dilution 10^–4, that of P. malariae at dilution 10^–7. Other Plasmodium species/strains were not tested due to lack of sample.

The melting curves of the created mixed Plasmodium infections at parasitaemia levels ranging from 0.1 to 0.4% (P. falciparum with either P. vivax, P. ovale wallikeri, P. ovale curtisi, P. knowlesi ATCC 30158, or P. cynomolgi KJ569868.1) are shown in Fig. 2. All tested mixed Plasmodium infections could be visually identified by the presence of more than one melting peak in the Texas Red and/or Cy5 channels, shifted melting temperatures, a shoulder added to the original melting peak, or the presence of a Cy5.5 melting peak.

The melting curves of the created mixed Plasmodium infections using dilution 1:128 of P. falciparum (initial parasitaemia level of 1.3%) with either P. vivax, P. ovale wallikeri, P. ovale curtisi, P. knowlesi ATCC 30158, or P. cynomolgi KJ569868.1 are shown in Additional file 2: Figure S2. Mixed infections with P. vivax, P. ovale wallikeri, and P. ovale curtisi could be visually identified. Mixed infections with P. malariae, P. knowlesi ATCC 30158, and P. cynomolgi KJ569868.1 could be readily identified up to dilution 1:64. In these mixed infections, P. falciparum could just be detected at dilution 1:128 by a small melting peak of the Cy5-labelled probe.

The 30 Cq-values used to construct the calibration curve are shown in Additional file 3: Table S1. The calibration curve itself is shown in Fig. 3. The exponential line of best fit has the equation: parasitaemia equals 57,012 * e^{(− 0.615*Cq-value)}. Figure 3 illustrates that variation in Cq-values in the lower range has a greater impact than variation in Cq-values in the higher range with regard to the calculated parasitaemia. For example, the difference between the calculated parasitaemia from a Cq-value of 15 (5.6%) and 16 (3.0%) is 2.6%, while the difference between the calculated parasitaemia from a Cq-value of 20 (0.26%) and 21 (0.14%) is 0.12%. The most commonly used cut-off in clinical practice is a parasitaemia level of 2%, which corresponds to a Cq-value of 16.68.

The results of the within-run and between-run precisions are summarized in Table 5. The within-run and between-run precisions at level 1 (average parasitaemia of 27.27%) were 4.6% and 7.8% CV, respectively. The between-run precisions equaled the within-run precision at level 3, 4 and 5 (average parasitaemia of 2.15%, 0.16% and 0.09%, respectively). This indicates stability under changing conditions.

No cross-reactivity was observed in the MC004 assay between the evaluated Plasmodium and non-Plasmodium species. The MC004 assay has been designed to specifically amplify Plasmodium DNA, and includes post-PCR melting curve analysis based on three different molecular beacon probes to determine specificity of the amplicons. Therefore, the occurrence of false-positive results in case of non-Plasmodium infections is deemed to be a highly unlikely possibility. This presumption is supported by the fact that cross-reactivity issues between Plasmodium and non-Plasmodium species have not been raised as a concern in previously published studies. However, cross-reactivity between different Plasmodium species is a well-known issue. In particular, the nucleotide sequences of P. vivax and P. knowlesi share high similarity, preventing specific PCR amplification in some assays. The same holds true for P. ovale curtisi and P. ovale wallikeri. This study shows these issues of cross-reactivity have been resolved by the MC004 assay. Arguably, cross-reactivity is the most likely explanation for the one discrepant result found with the MC004 assay (See Table 1, P. vivax (SKML sample 2018.1G). The result of MC004 was P. knowlesi, which was confirmed by Sanger sequencing), but it may also have been a clerical error or an oversight. The differentiation of P. knowlesi ATCC 30158 and P. knowlesi LT48 was verified, but no sample was available that contained P. knowlesi ATCC 30153. The only available P. cynomolgi sample showed melting temperatures consistent with strain KJ569868.1. Each Plasmodium species could be straightforwardly identified using the melting temperatures shown in Table 3. The MC004 assay cannot differentiate between P. vivax and P. simium, as their mitochondrial genomes differ at two nucleotides positioned outside the regions covered by the probes. Also, the genomes of P. malariae and P. brasilianum are very similar, and will have identical melting temperatures [22]. In clinical practice, this would not affect patient management. To reduce the probability of human error and to increase the speed of reporting results, species identification could be performed and reported automatically by programming the laboratory information management system, or semi-automatically by programming software like Microsoft Excel to return the species after entering the melting points.

The MC004 assay was capable of detecting 1 × 10^–3 IU/mL of the 1st WHO International Standard for P. falciparum DNA (NIBSC code: 04/176), which is equivalent to a dilution of 10^–12. The dilution series of two P. falciparum-infected whole blood samples (parasitaemia of 0.2% and 1.3%) shows that the limit of detection determined this way depends on the starting concentration (Table 4). Therefore, the detection limit was not expressed as number of parasites per volume of blood, but rather as the highest dilution of an arbitrary, relatively low, undiluted parasitaemia at which the Plasmodium species was consistently detected. Since there is no WHO standard available for other Plasmodium species than P. falciparum, a dilution series of clinical samples was used to assess the limit of detection for other Plasmodium species. Both the undiluted parasitaemia, ranging from < 0.1% to 0.3%, and the highest dilutions at which Plasmodium DNA was still detected, ranging from 10^–4 to 10^–7, were roughly comparable between P. falciparum, P. vivax, P. malariae, P. ovale curtisi and P. knowlesi ATCC 30158. No clinical sample was available to assess the limit of detection of P. cynomolgi. These limits of detection are regarded as sufficient to warrant further testing in both clinical and research settings.

Infections with more than one Plasmodium species are relatively rare in clinical practice, but relevant. Due to lack of sample, there was no opportunity to test actual samples from patients, besides the UK NEQAS sample (P. falciparum + P. vivax, See Table 1). Thus, mixed infections were mimicked by mixing DNA-isolates of P. falciparum and either P. vivax, P. ovale wallikeri, P. ovale curtisi, P. knowlesi ATCC 30158, or P. cynomolgi KJ569868.1. To further demonstrate the proof of principle that the MC004 assay can detect mixed infections, the amount of DNA of each Plasmodium species was kept roughly comparable, which was in the range of 0.1 to 0.4% parasitaemia (Fig. 2). An additional experiment, in which P. falciparum (with initial parasitaemia of 1.3%) was diluted up to 1:128, indicated that mixed infections may be missed if differences in DNA concentration between Plasmodium species are large (Additional file 2: Figure S2). P. falciparum is the most commonly identified species in our geographic area, and poses the greatest threat globally. Therefore, the focus was on mixed infections with P. falciparum, although any combination of two or even three Plasmodium species is theoretically possible. To what extent mixed infections are still detectable in cases where the difference between the Plasmodium species in parasitaemia are larger, or in case of infections with three Plasmodium species, remains to be verified by further studies, preferably empirically in the field. Features of the MC004 assay that specifically help to distinguish mixed infections with P. falciparum are the relatively large differences in melting temperatures of the Texas red and Cy5 labelled probes between P. falciparum and the other Plasmodium species, as well as the absence of a Cy5.5 melting temperature, which is only present in case of P. vivax, P. knowlesi or P. cynomolgi infections (Table 3).

One of the medical decision points to identify patients with severe or complicated malaria, particularly P. falciparum malaria, is more than two percent of the red blood cells parasitized [23, 24]. This arguably conservative threshold has traditionally been estimated by microscopy and is used rather than the WHO criterion of a parasitaemia estimate of greater than 4% [25]. A means to use the MC004 assay to estimate the level of parasitaemia based on the correlation between Cq-value and amount of Plasmodium DNA was sought out. The resulting calibration curve is shown in Fig. 3, which is traceable to by microscopy estimated levels of parasitaemia. To spare clinical sample and for sake of standardization, the WHO International Standard for P. falciparum DNA could be used to reestablish or verify the calibration curve when appropriate by assigning parasitaemia values to Cq-values. The performance of the MC004 assay when put into practice to estimate the parasitaemia of both P. falciparum and non-falciparum species at various levels and at different developmental stages, is to be shown empirically.

The within-run and between-run precisions close to the most relevant medical decision point of 2% parasitaemia (2.15%) and in the lower range (0.16% and 0.09%) and high range (27.27%) of less than 20% CV indicates robustness of the MC004 assay. While consistent results were obtained between runs and instruments, only one technician performed the testing, and one reagent lot MC004 was used. Nevertheless, considering the generally large variation in parasitaemia estimates by microscopy, precision is expected to improve with the MC004 assay.

This PCR protocol has limitations. First, differentiation between P. vivax and P. simium and between P. malariae and P. brasilianum is not possible, as the respective melting curves would be identical. Second, the detection of minority parasite populations in mixed infections may be compromised, if the differences in DNA concentration between Plasmodium species are too large. A methodological limitation of this study was the lack of field-collected parasites of each Plasmodium species or strain the PCR was designed to detect and identify, and of field-collected mixed infections.